# Ultron — full LLM corpus AI agents that run sales, marketing, and engineering work for founders. This file inlines Ultron's canonical product summary and every published skill so an LLM assistant has the full picture in one fetch. Canonical URLs: - Home: https://app.51ultron.com/ - BCP: https://app.51ultron.com/bcp - Skills catalogue: https://app.51ultron.com/docs/skills/catalog - OpenAPI: https://app.51ultron.com/openapi.json - Agent Skills index: https://app.51ultron.com/.well-known/agent-skills/index.json --- # What Ultron is Ultron is an AI operating system for founder-led companies. The product ships seven named agents — Cortex (intelligence), Specter (acquisition), Striker (growth), Pulse (marketing), Sentinel (engineering), Amplify (paid media), and Counsel (legal) — that share a discovery layer, a background-job runtime, and a sandboxed shell. Users invoke skills via slash commands or natural language; the discover_capability tool routes the request to the right skill and the chat-v2 SSE pipeline runs it. Architecture in one paragraph: a Next.js app hosts the chat UI and the SSE turn runner; missions and skills live in Supabase; long-running work is dispatched to Cloudflare Workers + Workflows (ultron-shell for code, ultron-capture for site cloning, orchestrator for scrape-and-rank jobs, marketing-swarm for carousel/video generation, discovery for the trend feed); Realtime triggers wake the chat session when background jobs settle. --- # Skills catalogue ## Competitive Analysis > Compare your product against named rivals — positioning, pricing, weaknesses, sales talking points. - URL: https://app.51ultron.com/docs/skills/competitive-analysis - Kind: Skill - Category: research - Owner: Cortex (/cortex) Produces a battlecard-style competitive analysis. For each rival: positioning vs you, pricing tiers, strengths and weaknesses, common buyer objections, and tactical talking points. Aggregates public web, marketing pages, Reddit threads, and prior research saved in memory. **When to use this** - user wants to compare their product against named competitors - user is preparing for a competitive deal, RFP, or bake-off - user asks for a battlecard, win/loss framing, or 'why pick us' talking points - user names two or more products and asks how they differ - user is positioning a new launch against incumbents **When NOT to use this** - user wants a single-company research report → use company-deep-dive - user wants to monitor competitor moves over time → use signal-news or signal-competitor-reviews - user wants pricing benchmarking only → use pricing-research - user wants tech-stack detection on a competitor → use signal-tech-stack **Example prompts** - `compare us to HubSpot and Salesforce` - `build me a battlecard against Apollo` - `how do we win against Outreach and Salesloft` - `what's our positioning vs Clay` - `us vs Lemlist — why pick us` - `competitive landscape for the cold-email space` **Inputs** - **competitors** — list of named competitor products to compare against - **angle** — optional: positioning, pricing, features, weaknesses, or 'all' **Output:** Markdown battlecard: positioning matrix, pricing comparison, per-competitor strengths/weaknesses, sales talking points. --- ## Company Deep Dive > Full dossier on one company — funding, leadership, tech stack, recent news. - URL: https://app.51ultron.com/docs/skills/company-deep-dive - Kind: Skill - Category: research - Owner: Cortex (/cortex) Thorough single-company research report. Pulls funding history, leadership team, tech stack hints, recent news, customer logos where visible, and any prior research saved in memory. Designed for due diligence, account research before a meeting, or 'should we partner with X' analysis. **When to use this** - user names ONE company and wants a thorough report - user asks for a dossier, deep dive, due diligence, or company profile - user is researching a target account before reaching out - user is evaluating a partner, vendor, or potential employer - user mentions a company and asks 'tell me everything about them' **When NOT to use this** - user names multiple companies for comparison → use competitive-analysis - user wants prep for a specific meeting with a person → use meeting-prep - user wants to track ongoing news about the company → use signal-news - user is evaluating ICP fit for outbound → use icp-deep-match **Example prompts** - `deep dive on Stripe` - `research ACME Corp for me` - `tell me everything about Vercel` - `company profile for Linear` - `due diligence on Notion` - `what should I know about Anthropic before the meeting` **Inputs** - **company_name** — single target company to research - **domain** — optional company domain to disambiguate - **angle** — optional focus: funding, team, product, tech-stack, customers, news, or 'all' **Output:** Structured dossier: company overview, funding history, leadership, tech stack hints, recent news, notable customers, risks. --- ## Pricing Research > Benchmark competitor pricing — tiers, anchors, packaging patterns. - URL: https://app.51ultron.com/docs/skills/pricing-research - Kind: Skill - Category: research - Owner: Cortex (/cortex) Pulls public pricing data for named competitors and benchmarks the user's pricing against the set. Identifies tier anchors, common packaging patterns, hidden gotchas (annual-only discounts, seat minimums), and where the user is over/underpriced. **When to use this** - user wants to compare competitor pricing - user is rebuilding their pricing page or considering a price change - user asks 'how do they price X' or 'what's the market rate for Y' - user mentions 'pricing benchmark' or wants tier-by-tier comparisons - user is questioning whether they're over- or underpriced **When NOT to use this** - user wants full competitive analysis with positioning + features → use competitive-analysis - user wants pricing-page COPY (not research) → use ads-copy or content-pulse - user wants to track pricing CHANGES over time → use signal-news **Example prompts** - `what does HubSpot charge for the pro tier` - `benchmark my pricing against the market` - `compare Salesforce and HubSpot pricing` - `what's the going rate for cold-email tools` - `are we underpriced vs Apollo` - `build me a pricing comparison table` **Inputs** - **competitors** — list of competitor names to benchmark - **tier_focus** — optional: 'starter', 'pro', 'enterprise', or 'all' **Output:** Pricing comparison table with tier anchors, packaging patterns, and a summary of where the user sits in the market. --- ## Meeting Prep > Brief on a specific upcoming meeting — attendee, company, talking points, hooks. - URL: https://app.51ultron.com/docs/skills/meeting-prep - Kind: Skill - Category: research - Owner: Cortex (/cortex) Targeted prep for a single upcoming meeting. Pulls the attendee's role and recent activity, the company's situation, recent news, prior touchpoints in the user's CRM/memory, and produces specific talking points and hooks tied to the meeting's purpose. **When to use this** - user has a meeting on the calendar and needs a brief - user mentions 'I have a call with [person] tomorrow' - user wants talking points or warm-up hooks for a known attendee - user is prepping for a discovery call, demo, or check-in - user wants to know what to ask or watch for in the meeting **When NOT to use this** - user is preparing a sales call deliverable (lead magnet / blueprint) → use striker-pre-call - user wants a generic company profile, not meeting-specific → use company-deep-dive - user wants to find PEOPLE at a target company (no meeting set) → use decision-maker-prospector - user wants to script the call itself → use objection-handler or striker-pre-call **Example prompts** - `prep me for the call with Jane at Acme tomorrow` - `I have a meeting with the CRO of Stripe — brief me` - `what should I know before talking to John` - `talking points for my discovery call with Linear` - `hooks for my call with the head of marketing at Notion` **Inputs** - **attendee_name** — name of the person being met - **attendee_company** — their company - **meeting_purpose** — discovery, demo, follow-up, partnership, etc. - **meeting_time** — optional date/time for context **Output:** Briefing card: who they are, company snapshot, recent news, hooks tied to meeting purpose, recommended questions, what to watch for. --- ## Funding Signals > Detect funding-round signals on watchlist accounts (free sources: News + EDGAR). - URL: https://app.51ultron.com/docs/skills/signal-funding - Kind: Skill - Category: research - Owner: Cortex (/cortex) Sweeps Google News RSS, Brave web search, and SEC EDGAR Form D for funding-round events on a list of target companies. Each detected raise tagged with amount, lead investor, date, and a recency tier (Hot 0-2w / Warm 2-4w / Cool 4-8w). Cross-checks two sources before confirming. **When to use this** - user wants to detect which target accounts just raised - user mentions 'who got funded recently' or 'recent raises in [space]' - user wants to time outreach to the post-funding window - user has a watchlist and wants to know which raised in the last N weeks - user is looking for net-new funded companies in a vertical **When NOT to use this** - user wants to find ALL VCs/angels for THEIR raise → use vc-prospector - user wants generic company news, not just funding → use signal-news - user wants to combine ALL signal types into one feed → use signal-multi-aggregator **Example prompts** - `which of my target accounts just raised` - `find recent Series A raises in fintech` - `who got funded in the last 4 weeks in my watchlist` - `post-funding signals on my account list` - `fresh raises in the AI space this month` **Inputs** - **watchlist** — list of company names or domains to scan - **window_days** — optional lookback window (default 60) **Output:** Per-company entries with amount, lead investor, date, source URLs, and recency tier (Hot/Warm/Cool). --- ## Company News Signals > Detect company events on watchlist accounts — leadership change, M&A, launches, layoffs, lawsuits. - URL: https://app.51ultron.com/docs/skills/signal-news - Kind: Skill - Category: research - Owner: Cortex (/cortex) Pulls Google News RSS + Brave search for event signals on a watchlist of target companies. Categorizes by type (leadership change, M&A, product launch, layoffs, customer wins, lawsuit) and polarity. Negative events (layoffs, lawsuits) trigger a 60-day backoff flag for outreach skills. **When to use this** - user wants ongoing news monitoring on target accounts - user mentions 'what's happening at [company]' or 'recent news on my watchlist' - user wants leadership-change alerts (great outreach trigger) - user is tracking accounts for product launches, M&A, customer wins - user wants to know which accounts are in trouble (layoffs, lawsuits) so they back off **When NOT to use this** - user wants funding signals specifically → use signal-funding - user wants competitor reviews / pain themes → use signal-competitor-reviews - user wants macro tech trends, not account-specific → use trend-feed - user wants a one-time deep dive on ONE company → use company-deep-dive **Example prompts** - `what's happening at my watchlist accounts this week` - `any leadership changes at the companies I track` - `news monitor on my target list` - `did anyone in my account list announce layoffs` - `M&A activity on my watchlist` **Inputs** - **watchlist** — list of company names or domains to scan - **window_days** — optional lookback window (default 30) - **event_types** — optional filter: leadership, launch, M&A, layoffs, lawsuit, customer-win **Output:** Grouped events per company tagged with type, polarity, source URL, date, and outreach-window flag. --- ## Tech Stack Signals > Detect what tech a target runs — analytics, CRM, payments, automation — and where the gaps are. - URL: https://app.51ultron.com/docs/skills/signal-tech-stack - Kind: Skill - Category: research - Owner: Cortex (/cortex) Direct site scrape + HTML signature matching against ~30 well-known tools (HubSpot, Segment, Mixpanel, Stripe, Auth0, Intercom, etc.). Surfaces stack GAPS that match what the user sells — those become outreach hooks. Free-only, no BuiltWith / Wappalyzer paid lookups. **When to use this** - user wants to know what tech a target company runs - user is qualifying for outbound based on tech-stack fit - user mentions 'do they use [tool]' or 'what's their stack' - user wants to find prospects MISSING a tool the user sells - user is building tech-stack-based ICP segments **When NOT to use this** - user wants paid stack data (BuiltWith, Wappalyzer) → not supported here - user wants company news, not stack → use signal-news - user wants to assess overall company fit (not just stack) → use icp-deep-match **Example prompts** - `does Acme use HubSpot` - `what's the tech stack at Stripe` - `which of my targets are missing analytics` - `find companies on my watchlist using Salesforce` - `tech stack signals for my account list` **Inputs** - **watchlist** — list of domains to fingerprint - **tools_of_interest** — optional list of specific tools to check for or detect gaps in **Output:** Per-domain stack readout: detected tools, confidence per detection, and gap analysis vs the user's product. --- ## Competitor Review Signals > Find buying-intent in competitor pain — bad G2/Capterra reviews, Reddit switching threads. - URL: https://app.51ultron.com/docs/skills/signal-competitor-reviews - Kind: Skill - Category: research - Owner: Cortex (/cortex) Scrapes G2 / Capterra reviews and Reddit complaint threads on competitor products. Returns identifiable reviewers (where shown) plus aggregate pain themes. The reviewers are warm prospects; the pain themes feed positioning and content. **When to use this** - user wants to find unhappy users of a competitor - user mentions 'who's complaining about [competitor]' or 'switching threads' - user wants pain-theme research for positioning or content - user is hunting for warm prospects from G2/Capterra/Reddit - user wants to harvest review-based objections **When NOT to use this** - user wants positioning vs a competitor (no reviews) → use competitive-analysis - user wants tech-stack-based intent → use signal-tech-stack - user wants funding/news monitoring → use signal-funding or signal-news **Example prompts** - `who's complaining about Apollo on G2` - `Reddit threads about people leaving Outreach` - `negative Capterra reviews of Salesloft` - `pain themes for HubSpot users` - `find warm prospects from competitor reviews` **Inputs** - **competitors** — list of competitor product names - **pain_focus** — optional: specific pain points to search for **Output:** Reviewer list (with identifying info where present) + aggregated pain themes + sample quotes for content. --- ## Buying Intent Aggregator > Combine all detected signals into one ranked buying-intent feed for outreach prioritization. - URL: https://app.51ultron.com/docs/skills/signal-multi-aggregator - Kind: Skill - Category: research - Owner: Cortex (/cortex) Reads memories saved by the other signal skills (funding / news / tech-stack / competitor-reviews) and applies tier-based scoring with recency multipliers and a cross-source bonus. Writes to the account_signals table — that table is what feeds the user's outreach prioritization. **When to use this** - user wants ONE ranked list of accounts to reach out to NOW - user mentions 'which accounts should I prioritize' or 'top buying-intent' - user has run multiple signal skills and wants them combined - user wants a daily/weekly intent digest across their watchlist **When NOT to use this** - no signal data has been gathered yet → run signal-funding / signal-news / etc. first - user wants ONE specific signal type → use that signal skill directly - user wants account research, not prioritization → use company-deep-dive **Example prompts** - `which accounts have the strongest buying intent right now` - `ranked outreach priority list` - `combine all my signals into one feed` - `top 20 accounts to hit this week` - `intent digest across my watchlist` **Inputs** - **watchlist** — optional — defaults to all accounts the user has signal data on - **limit** — optional cap on the ranked list (default 20) **Output:** Ranked feed of accounts with composite score, contributing signals, recency, and recommended outreach action. --- ## Trend Feed > Macro-trend digest — what's hot in AI/tech this week across HN, Reddit, GitHub, PH, HF Papers. - URL: https://app.51ultron.com/docs/skills/trend-feed - Kind: Skill - Category: research - Owner: Cortex (/cortex) Aggregates Hacker News, Reddit, GitHub Trending, HuggingFace Papers, and Product Hunt into a weekly macro-trend digest. Different from signal-news (account-specific) — this is market-level. Velocity tracking via memory: tracks what's accelerating vs what's fading. **When to use this** - user wants a digest of what's trending in tech / AI / their space - user mentions 'what's hot this week', 'trends', or 'macro' - user wants content fuel for thought-leadership posts - user is hunting for early-mover signals on new tools or topics - user wants to track which AI papers / repos / launches are accelerating **When NOT to use this** - user wants news on a specific company → use signal-news - user wants signals on their watchlist → use signal-multi-aggregator - user wants a deep dive on one trend → use company-deep-dive or content-pulse **Example prompts** - `what's trending in AI this week` - `macro-trend digest` - `what's hot on HN and Product Hunt` - `AI papers that are accelerating` - `give me content fuel from this week's tech trends` **Inputs** - **vertical** — optional: AI, dev-tools, sales-tech, etc. - **window_days** — optional lookback (default 7) **Output:** Weekly digest with categorized trends, velocity tags (accelerating/steady/fading), and top items per source. --- ## Discovery Lab Research > Search Ultron's curated Discovery Lab library — playbooks, viral formats, trends, funding, and live feed items. - URL: https://app.51ultron.com/docs/skills/discovery-lab-research - Kind: Skill - Category: research - Owner: Cortex (/cortex) Queries Ultron's Discovery Lab Vectorize index (long-form playbooks + Exploding Topics + live worker feeds) and returns ranked passages with title, summary, preview image, and link. Does NOT scrape the web — only returns what's already indexed. **When to use this** - user wants to search the library / Discovery Lab - user mentions 'what does the library say about X' / 'find playbooks on X' / 'pull library examples' - user wants cached viral format / playbook / trend writeups - user wants curated examples or case studies on a topic - user asks about Exploding Topics insights - user wants a quick scan of cached signals (funding rounds, GitHub trending, product launches) on a theme **When NOT to use this** - user wants live web research → use signal-funding / signal-news / signal-tech-stack - user wants a full company dossier → use company-deep-dive - user wants to compare competitors → use competitive-analysis - user wants to write new content → use content-pulse **Example prompts** - `search the library for going viral` - `what does our library say about AI agent pricing` - `find playbooks on cold-email at scale` - `pull library examples on dev-tool launches` - `discovery lab search: TikTok hooks` - `show me cached funding rounds in dev tools` **Inputs** - **query** — natural-language query, the more specific the better - **kind** — optional filter: library | topic | startup | insight | feed - **limit** — optional max results (default 10) **Output:** Ranked passages with title, summary, preview image, source link, and kind tag. One-paragraph synthesis at the end. --- ## Google Maps Leads > Scrape brick-and-mortar businesses from Google Maps — dental clinics, gyms, restaurants, agencies. - URL: https://app.51ultron.com/docs/skills/gmaps-leads - Kind: Skill - Category: lead-gen - Owner: Specter (/specter) Background Apify job (ultron-gmaps actor): runs a Google Maps search for a business type in a location and extracts each place's name, address, phone, website, rating, hours, categories. Results stream into the leads table over 2-3 min per ~100 places. **When to use this** - user wants local-business leads (brick-and-mortar) - user names a business type and a location ('dentists in Lisbon', 'gyms in Madrid') - user mentions 'Google Maps' / 'scrape Maps' / 'list of [businesses] in [city]' - user is doing local-business outbound (services, agencies, contractors) - user wants leads with phone + website (not just emails) **When NOT to use this** - user wants B2B SaaS leads / decision-makers → use decision-maker-prospector - user wants investors → use vc-prospector - user wants people who hold a specific role → use decision-maker-prospector or hiring-manager-prospector - user wants a single-business deep dive → use company-deep-dive **Example prompts** - `find me dentists in Lisbon` - `list of gyms in Madrid` - `scrape Google Maps for restaurants in Brooklyn` - `agencies in Berlin` - `100 dental clinics in Bucharest` - `give me a lead list of contractors in Austin` **Inputs** - **search_terms** — list of business types ('dental clinic', 'orthodontist') - **location** — city or geography for the search - **max_per_search** — optional cap per term (default 100) - **include_reviews** — optional, default false (slower if true) **Output:** Background job. Places stream into the leads table with name, address, phone, website, rating, categories. ETA 2-3 min per 100. --- ## First Cold Email > Write the FIRST cold email — opener, body, single CTA. Seniority-aware (ATL strategic / BTL operational). - URL: https://app.51ultron.com/docs/skills/email-first-touch - Kind: Skill - Category: lead-gen - Owner: Specter (/specter) Generates a first-touch cold email. Takes a seniority parameter — ATL (VP / C-Level) gets 2-3 sentences, strategic, outcome-oriented; BTL (Manager / IC) gets 4-6 sentences, operational, day-in-the-life. Picks one of 4 opener formulas. Single CTA discipline. Hard banlist for AI tells (leverage / unlock / transform / etc). **When to use this** - user wants to write the first cold email to a prospect - user mentions 'cold email', 'first touch', 'opening message', 'email 1' - user wants opener language for outreach - user is starting a new sequence and needs the entry email **When NOT to use this** - user wants follow-ups (emails 2, 3, breakup) → use email-follow-up-cadence - user wants to revive a ghosted/lost lead → use email-re-engagement - user wants post-meeting follow-up → use striker-follow-up - user wants to personalize a TEMPLATE across MANY leads → use email-personalize-batch - user wants subject lines only → use email-subject-lines **Example prompts** - `write a cold email to the VP of Sales at Acme` - `first touch for John, Head of Marketing at Stripe` - `draft email 1 of my sequence` - `cold opener for a CFO` - `email a Series B SaaS founder` - `SDR email for our outbound push` - `BDR email to a CTO at a Series B company` - `outbound email to a Head of RevOps` - `prospecting email for our cold push` **Inputs** - **prospect** — name + role + company of the recipient (or lead_id from CRM) - **seniority** — atl, btl, or auto - **angle** — optional hook angle the user wants emphasized **Output:** One first-touch email: subject, opener, body, CTA, plus a one-line rationale. --- ## Follow-Up Cadence > COLD sequence — emails 2, 3, and breakup. The prospect never replied to email 1. They don't know you yet. - URL: https://app.51ultron.com/docs/skills/email-follow-up-cadence - Kind: Skill - Category: lead-gen - Owner: Specter (/specter) Produces the rest of a COLD-OUTREACH sequence after email 1 went unanswered. Each follow-up uses a DIFFERENT angle from email 1 (reframe / proof point / resource drop / gentle disqualifier / question / walkaway) and is shorter than the previous. Default cadence +3 → +5 → +7 business days. The prospect has not engaged yet. **When to use this** - user wants emails 2, 3, breakup of a COLD sequence - user mentions 'cold follow-up', 'they never replied', 'no response to email 1', 'breakup email', 'bump' - user has email 1 written and wants the rest of the cold cadence - user wants a 3-step or 4-step cold-sequence push **When NOT to use this** - user wants email 1 (first cold touch) → use email-first-touch - user wants ONE single email after a real interaction (meeting/call/demo) → use post-meeting-followup (id: follow-up) - user wants an IN-DEAL multi-step sequence after a sales call → use striker-follow-up - user wants to revive a ghosted/lost lead from MONTHS ago who already KNEW them → use email-re-engagement - user wants subject lines only → use email-subject-lines **Example prompts** - `they never replied to email 1 — write 2 and 3` - `cold sequence follow-ups` - `breakup email for cold prospects who haven't replied` - `no response — what's email 2` - `draft a 4-step cold cadence` **Inputs** - **email_1_text** — the first-touch email to chain off of - **cadence** — optional spacing in business days (default +3 +5 +7) - **breakup** — optional flag to include final walkaway email **Output:** Emails 2, 3, and (optional) breakup with subjects, bodies, and rationale per step. --- ## Email Re-Engagement > Win-back ghosted, lost, or churned prospects with a real 'what's new' angle. - URL: https://app.51ultron.com/docs/skills/email-re-engagement - Kind: Skill - Category: lead-gen - Owner: Specter (/specter) For prospects who already KNEW you — ghosted after a demo, lost to a competitor, churned customer, fizzled-warm prospect. Diagnoses the cool type, requires a real 'what's new' angle, and uses a low-pressure tone. NOT for cold first-touch — these prospects already have context. **When to use this** - user wants to revive a lost or ghosted lead - user mentions 're-engage', 'win-back', 'wake up', 'lost deal revival', 'churned customer' - user has prospects from months/years ago they want to reach back out to - user wants to follow up on a deal that fizzled after demo **When NOT to use this** - prospect has never heard of the user → use email-first-touch - user is in an active sequence with no reply yet → use email-follow-up-cadence - user wants post-meeting follow-up → use striker-follow-up **Example prompts** - `win-back email for a customer who churned` - `re-engage prospects who ghosted after demo` - `wake up lost deals from Q2` - `what's new email for old leads` - `revive my lost-to-competitor list` **Inputs** - **cool_type** — ghosted-after-demo, lost-to-competitor, churned, fizzled-warm, never-converted - **whats_new** — the real reason to reach out now — new product, new ICP fit, etc. - **prospect** — lead id or name + company **Output:** One re-engagement email tuned to the cool type, plus a follow-up if no reply. --- ## Email Subject Lines > 5-8 subject-line variants tagged with formula, character count, and rationale. - URL: https://app.51ultron.com/docs/skills/email-subject-lines - Kind: Skill - Category: lead-gen - Owner: Specter (/specter) Shared utility — Specter and Striker both call this. Returns 5-8 variants per request, tagged with formula (specific outcome, curiosity gap, their language, pattern interrupt, trigger, direct ask) and character count and rationale. Used as a sub-step of other email skills or directly when the user wants subject options. **When to use this** - user wants subject-line variants for any email type - user mentions 'subject line', 'subject', 'A/B test subjects' - user is unhappy with their current subject and wants alternatives - user wants short / curiosity-gap / specific-outcome subject options **When NOT to use this** - user wants the full email body — use email-first-touch / striker-follow-up / etc - user wants ad headlines, not email subjects → use ads-copy - user wants social hooks, not subjects → use content-hooks **Example prompts** - `subject line ideas for this cold email` - `give me 8 subject options` - `A/B test subjects for my sequence` - `shorter subject for my follow-up` - `curiosity-gap subject lines` **Inputs** - **email_body** — the email the subject should match - **tone** — optional — formal, casual, curious, direct - **variants** — optional count (default 6) **Output:** 5-8 subject variants with formula tag, character count, and one-line rationale. --- ## Batch Email Personalize > Apply ONE template across MANY leads with real per-lead specificity. Cap 30 per batch. - URL: https://app.51ultron.com/docs/skills/email-personalize-batch - Kind: Skill - Category: lead-gen - Owner: Specter (/specter) Takes a template + a list of leads (CSV / lead_ids / inline list) and produces personalized variants per lead. Hard cap of 2 enrichment calls per lead. Tags each output strong / acceptable / fallback so the user knows which to keep. Flags batches with >40% fallback as a warning that the template doesn't fit the audience. **When to use this** - user has a TEMPLATE and a LIST and wants per-lead personalization - user mentions 'personalize this for my list', 'batch personalization' - user has 5-30 leads and wants each tailored without writing each one manually - user is sending a sequence to a target list and wants real specificity **When NOT to use this** - user wants ONE email to ONE prospect → use email-first-touch - user wants a generic template (not personalized) → use email-first-touch - user has more than 30 leads → run multiple batches - user wants to find leads first → use decision-maker-prospector **Example prompts** - `personalize this email for my 20 leads` - `batch personalize this template for the list` - `I have 30 prospects and one template — tailor each` - `personalize my sequence across this CSV` **Inputs** - **template** — the email template with merge points - **leads** — list of lead ids, CSV, or inline list (max 30) - **enrichment_budget** — optional, default 2 calls per lead **Output:** Personalized email per lead, tagged strong / acceptable / fallback, plus a batch quality summary. --- ## Lead Scoring > Score and prioritize existing leads on fit + engagement signals. Writes scores back to the CRM. - URL: https://app.51ultron.com/docs/skills/lead-scoring - Kind: Skill - Category: lead-gen - Owner: Specter (/specter) Reads the user's existing leads, scores each on a fit dimension (ICP match) and an engagement dimension (recency of interaction), and writes the composite score back to the lead row. Surfaces the top N for outreach. **When to use this** - user wants to prioritize the leads they already have - user mentions 'score my leads', 'who should I reach out to first', 'ranking my list' - user has a long list and wants to focus on the top 20 - user wants engagement-based prioritization **When NOT to use this** - user wants to FIND new leads → use decision-maker-prospector - user wants to evaluate ICP fit for ONE company → use icp-deep-match - user wants signals across accounts (not just leads) → use signal-multi-aggregator **Example prompts** - `score my leads` - `who should I reach out to first` - `rank my prospect list` - `top 20 leads by fit` - `prioritize my outreach list` **Inputs** - **filter** — optional filter for which leads to score (status, source, etc.) - **weights** — optional override for fit-vs-engagement weights **Output:** Updated lead rows with score + reasoning, plus a top-N list for immediate action. --- ## ICP Deep Match > Deep-evaluate whether a single company or lead matches the user's ICP — fit score + reasons. - URL: https://app.51ultron.com/docs/skills/icp-deep-match - Kind: Skill - Category: lead-gen - Owner: Specter (/specter) Runs a thorough ICP fit check on ONE target. Pulls company info (sector, headcount, signals), compares to the user's defined ICP, and returns a fit score with specific reasons. Useful before doing outbound to a single high-value account. **When to use this** - user wants to evaluate ONE company for ICP fit - user mentions 'is X a fit', 'should I target Y', 'ICP match' - user is qualifying a single account before investing time - user wants reasons (not just a score) for fit **When NOT to use this** - user has many leads to score → use lead-scoring - user wants to research the company generally → use company-deep-dive - user wants positioning/comparison → use competitive-analysis **Example prompts** - `is Stripe a fit for our ICP` - `ICP match score for ACME` - `should I target Notion` - `deep ICP check on Linear` - `is this account worth pursuing` **Inputs** - **target** — company name or domain to evaluate - **icp_definition** — optional override of the user's stored ICP **Output:** Fit score, dimension-by-dimension breakdown, specific reasons, and a 'pursue / pass / monitor' recommendation. --- ## Trigger Events > Find recent buying-trigger events (funding, hiring, launches) for target accounts. - URL: https://app.51ultron.com/docs/skills/trigger-events - Kind: Skill - Category: lead-gen - Owner: Specter (/specter) Sweeps target accounts for buying-trigger events: funding, hiring waves, product launches, leadership changes, expansions. Combines logic from signal-funding + signal-news into a one-shot 'why now' summary tied to outreach timing. **When to use this** - user wants 'why now' triggers for their account list - user mentions 'trigger events', 'why now', 'recent activity at my targets' - user is composing outreach and needs a hook tied to a recent event - user wants the freshest signals across an account list **When NOT to use this** - user wants ONLY funding signals → use signal-funding - user wants ONLY news monitoring → use signal-news - user wants the unified ranked feed → use signal-multi-aggregator **Example prompts** - `trigger events on my watchlist` - `why-now hooks for my accounts` - `recent activity at my target companies` - `find me hooks for outreach this week` - `fresh triggers in my pipeline` **Inputs** - **watchlist** — list of accounts to scan - **event_types** — optional filter — funding, hiring, launch, leadership, expansion - **window_days** — optional lookback (default 30) **Output:** Per-account trigger summary with type, date, source, and a suggested outreach hook. --- ## Cold Outreach Campaign > End-to-end cold campaign — research prospects + craft a personalized cold sequence. - URL: https://app.51ultron.com/docs/skills/specter-cold-outreach - Kind: Skill - Category: lead-gen - Owner: Specter (/specter) Multi-step Specter workflow: research the prospect set, segment by ICP, craft the first email plus 2-3 follow-ups, generate subject variants, optionally enrich emails. Triggered by /specter for cold-outreach work. NOT for fundraising — that's a separate flow. **When to use this** - user typed /specter and wants a cold campaign end-to-end - user wants research + sequence (not just one email) - user wants a coordinated outbound push across a target list - user mentions 'cold outreach campaign', 'outbound push', 'sequence + research' **When NOT to use this** - user wants ONE email only → use email-first-touch - user is fundraising (investors, not buyers) → use vc-prospector + fundraising_outreach - user wants to enrich an existing list (no campaign) → use specter-lead-research **Example prompts** - `/specter run a cold campaign to my Series B SaaS list` - `cold outreach push to dental clinics in Lisbon` - `outbound campaign for VPs of Engineering` - `build me an end-to-end cold campaign for fintechs` - `/specter prospect + sequence for our launch` **Inputs** - **target_segment** — ICP filter or list of accounts - **angle** — the outreach angle — pain point, trigger event, etc. - **sequence_length** — optional, default 3 emails **Output:** A research summary, ICP segmentation, full email sequence with subjects, and a list of personalized variants. --- ## LinkedIn Prospector > Pull a targeted lead list from LinkedIn: buyer decision-makers, hiring-signal buyers, or investors, streamed into the CRM as a background job. - URL: https://app.51ultron.com/docs/skills/linkedin-prospector - Kind: Skill - Category: lead-gen - Owner: Specter (/specter) One dispatcher for the three LinkedIn prospecting shapes. Buyer mode finds decision-makers by role and company filters; hiring-signal mode targets companies whose open roles reveal the buying problem; investor mode targets VCs and angels for a raise. Fires the LinkedIn Prospector actor in the background; leads stream into the CRM over a few minutes. **When to use this** - you want a list of decision-makers who buy what you sell - you want companies hiring for a role that signals your problem space - you are raising and want a targeted investor list **When NOT to use this** - writing the outreach itself: use specter-cold-outreach - local or SMB business lists: use gmaps-leads **Example prompts** - `Find 50 heads of RevOps at Series B SaaS companies` - `Which companies are hiring SDR managers right now?` - `Build me a seed-stage AI investor list` **Output:** A saved, deduped lead list in the CRM with role, company, and profile fields, plus a run summary. --- ## Post-Meeting Follow-Up > ONE email after a real interaction (meeting / call / demo). Recap + next steps. NOT a sequence, NOT cold. - URL: https://app.51ultron.com/docs/skills/follow-up - Kind: Skill - Category: sales - Owner: Striker (/striker) Single follow-up email sent AFTER the prospect already knows you — discovery call, demo, conference booth, intro meeting. Includes a recap of what was discussed, agreed next steps, and a clear CTA. One email, not a sequence. **When to use this** - user wants ONE email after a meeting, call, demo, or conference - user mentions 'thank you email', 'recap email', 'next steps', 'post-meeting', 'after the call' - user wants to lock in agreed next steps with a written recap - the prospect already knows the user — this is in-deal, not cold **When NOT to use this** - user wants a SEQUENCE (multiple emails over days) → use striker-follow-up for in-deal sequences, email-follow-up-cadence for cold sequences - user is following up on a COLD email that got no reply (prospect doesn't know them) → use email-follow-up-cadence - user wants to revive a deal from MONTHS ago that went cold → use email-re-engagement - user is in /striker post-call mode and wants a structured deliverable package → use striker-post-call **Example prompts** - `follow-up email after my demo with Acme` - `thank-you email for the meeting yesterday` - `send a recap to the team I just met` - `next-steps email after my discovery call` - `one quick follow-up after the conference chat` **Inputs** - **meeting_context** — what the meeting was about, key points discussed - **next_steps** — what was agreed - **tone** — optional — formal, casual **Output:** One follow-up email with recap, agreed next steps, and clear CTA. --- ## Objection Handler > Talking points for sales objections — too expensive, not the right time, need to think. - URL: https://app.51ultron.com/docs/skills/objection-handler - Kind: Skill - Category: sales - Owner: Striker (/striker) When a prospect pushes back, this skill produces specific responses tailored to the objection type. Covers price, timing, authority, fit, and 'I need to think.' Includes reframing language, validating phrases, and a recovery question to move the conversation forward. **When to use this** - user is facing a sales objection mid-conversation - user mentions 'objection', 'pushback', 'they said no', 'too expensive', 'not the right time', 'need to think' - user wants reframing language for a price / timing / fit objection - user is preparing for likely objections before a call **When NOT to use this** - user wants to research a competitor that triggered the objection → use competitive-analysis - user wants to revive a deal they've already lost → use email-re-engagement - user wants pricing benchmarking (not objection handling) → use pricing-research **Example prompts** - `they said we're too expensive — how do I respond` - `objection: not the right time` - `prospect pushed back on price` - `they said they need to think about it` - `talking points for the 'we already use [competitor]' objection` **Inputs** - **objection** — the objection language verbatim if possible - **context** — deal stage and any relevant prospect context **Output:** Recommended response language + reframe + recovery question to keep the deal alive. --- ## Pre-Call Briefing > Research the prospect for a sales call — produces lead magnet + sales blueprint deliverables. - URL: https://app.51ultron.com/docs/skills/striker-pre-call - Kind: Skill - Category: sales - Owner: Striker (/striker) Multi-step workflow: deep research the prospect, build a 'lead magnet' canvas (a value-first asset for the call), and a sales blueprint canvas (discovery + demo + objection map). Triggered by /striker pre-call. Output canvases get auto-built from skill-tagged JSON. **When to use this** - user typed /striker and is preparing for a sales call - user mentions 'pre-call', 'call prep', 'sales deck for the call', 'lead magnet' - user wants a value-first asset to bring to a discovery call - user wants a structured sales blueprint for the call **When NOT to use this** - user wants generic meeting prep (not deliverables) → use meeting-prep - user wants pure company research → use company-deep-dive - user is post-call and wants to send a package → use striker-post-call **Example prompts** - `/striker prep me for the call with Acme tomorrow` - `build a lead magnet for my call with Stripe` - `pre-call research and sales blueprint` - `I have a discovery call next week — Striker, do your thing` **Inputs** - **prospect_name** — person being met - **prospect_company** — their company - **meeting_purpose** — discovery, demo, expansion, partnership **Output:** Two canvases: lead-magnet (value-first asset) and sales-blueprint (discovery + demo + objection map). --- ## Post-Call Package > Build the post-call 'we already started' package — extra context, extras, recap deliverable. - URL: https://app.51ultron.com/docs/skills/striker-post-call - Kind: Skill - Category: sales - Owner: Striker (/striker) After a sales call: gathers extra context based on what came up in the call, identifies extras Striker can deliver (industry stats, competitive intel, customer logos, content the prospect would value), and produces a post-call canvas + email. The 'we already started building for you' deliverable. **When to use this** - user typed /striker after a sales call - user mentions 'post-call', 'after the call', 'call deliverable', 'follow-up package' - user wants the deliverable that lands AFTER discovery to differentiate - user wants to wow the prospect with proactive prep **When NOT to use this** - user wants a normal recap email → use follow-up - user wants an in-deal follow-up sequence (3 emails) → use striker-follow-up - user is preparing for a future call → use striker-pre-call **Example prompts** - `/striker build the post-call package for Acme` - `post-call deliverable after my discovery with Stripe` - `we already started building for you — package` - `wow my prospect with a post-call canvas` - `Striker, prep extras after my demo` **Inputs** - **prospect_name** — person from the call - **prospect_company** — their company - **call_notes** — what was discussed, pain points, asks **Output:** Post-call canvas with extras tied to the call + a recap email + suggested next steps. --- ## Content Pulse > Unified content engine — ingest sources, run a brief, produce 3 variants per format. - URL: https://app.51ultron.com/docs/skills/content-pulse - Kind: Skill - Category: content - Owner: Pulse (/pulse) The main content entry point. Takes references / swipe files / transcripts / 'I want to write about X', extracts anatomy, runs a brief interview, and produces 3 variations per requested format. Replaces the retired generate_hook + generate_caption tools. Composes other content skills (hooks / format-pick / brand-voice) under the hood. **When to use this** - user wants to generate content from sources or just an idea - user mentions 'write a post', 'help me write about', 'I want to write about [topic]', 'give me 3 variants of a post' - user has references and wants content built on top of them - user is unsure of format and wants the engine to decide + produce - user wants content variants (3 versions) of a piece they want to write about a topic or launch **When NOT to use this** - user wants ONE specific format and knows it (LinkedIn post / blog outline / social) → use the targeted skill - user wants to repurpose ONE canonical asset into many → use content-repurpose - user wants the output cleaned of AI tells AFTER generation → use humanizer - user wants a visual canvas (carousel / chart) → use canvas-intelligence **Example prompts** - `write a post about X based on this transcript` - `I want to write about our launch — give me 3 variants` - `turn this article into a LinkedIn post` - `content from my swipe file` - `help me write about Apify scrapers` - `write a polished caption for my carousel` - `polished caption for my post` - `social caption for the launch` **Inputs** - **sources** — URLs, transcripts, swipe-file refs (optional) - **topic** — the subject / angle - **formats** — optional list — linkedin, twitter, blog, etc. **Output:** 3 content variants per requested format with hooks, body, and a rationale per variant. --- ## Brand Voice > Define / refine / audit the user's brand voice — the constraints other content skills clamp to. - URL: https://app.51ultron.com/docs/skills/brand-voice - Kind: Skill - Category: content - Owner: Pulse (/pulse) Produces a structured Voice Profile: tone descriptors, sentence rhythm, banlist, signature phrases, examples. Saved to the user's content library. All other content skills read this profile to keep output on-brand. Run once when onboarding or whenever the voice drifts. **When to use this** - user wants to define their brand voice for the first time - user mentions 'set my tone', 'audit my voice', 'why does my content sound off', 'voice consistency' - user wants to update or refine an existing voice profile - user is onboarding to Pulse **When NOT to use this** - user wants to write content NOW (not define voice) → use content-pulse or a targeted skill - user wants to remove AI patterns from existing content → use humanizer **Example prompts** - `set my tone of voice` - `audit my LinkedIn for voice consistency` - `why does my content sound generic` - `build my voice profile from these 5 posts` **Inputs** - **sample_posts** — 5-10 of the user's existing posts (optional, for audit) - **desired_tone** — optional descriptors — formal, casual, sharp, warm **Output:** Structured Voice Profile (tone, rhythm, banlist, signature phrases, examples) saved to content library. --- ## Humanizer > Humanize a draft — strip AI tells, clean dashes, inject voice. Post-process content before sending. - URL: https://app.51ultron.com/docs/skills/humanizer - Kind: Skill - Category: content - Owner: Pulse (/pulse) Takes any content output and runs a polish pass: removes AI patterns (em-dashes, 'in conclusion', 'leverage', 'unlock', 'transform', 'in today's fast-paced world'), normalizes punctuation, and re-injects the user's brand voice. POST-PROCESSOR — should be called on the output of other content skills before showing the user. **When to use this** - user pastes content and asks to 'humanize', 'de-AI', 'make less robotic' - user mentions 'remove AI patterns', 'clean up dashes', 'edit for naturalness' - user wants a polish pass on existing content - any time output of another content skill is about to be shown — auto-chain **When NOT to use this** - user wants to GENERATE content, not polish → use content-pulse or targeted skills - user wants to define voice, not apply it → use brand-voice **Example prompts** - `humanize this` - `de-AI my post` - `polish this draft` - `remove the AI feel from this email` - `clean up the em-dashes` **Inputs** - **content** — the text to humanize - **apply_voice** — optional flag to also re-apply brand voice (default true) **Output:** Polished version of the input with AI tells removed and voice consistent. --- ## Content Hooks > 3-5 scroll-stopping hooks (first 1-2 sentences) tagged with formula + char count. - URL: https://app.51ultron.com/docs/skills/content-hooks - Kind: Skill - Category: content - Owner: Pulse (/pulse) Shared utility — content-pulse calls this internally. Produces 3-5 hook variants for a piece of content, each tagged with formula (specific outcome / curiosity gap / their language / pattern interrupt / contrarian / tension) and character count and rationale. **When to use this** - user wants better hooks for an existing piece - user mentions 'better hook', 'opening line', 'first 210 chars', 'scroll-stopper' - user is unhappy with their hook and wants alternatives - user wants A/B hook variants for the same idea **When NOT to use this** - user wants the full post body → use linkedin-post or content-pulse - user wants email subjects → use email-subject-lines - user wants ad headlines → use ads-copy **Example prompts** - `better hook for this post` - `scroll-stopping opener for my LinkedIn post` - `5 hook variants` - `first 210 chars for my post` **Inputs** - **post_body** — the content the hook should match - **platform** — optional — linkedin, twitter, etc. - **variants** — optional count (default 4) **Output:** 3-5 hook variants tagged with formula + character count + rationale. --- ## Content Format Picker > Pick the best format for an idea — text post vs carousel vs video vs poll — anchored to the goal. - URL: https://app.51ultron.com/docs/skills/content-format-pick - Kind: Skill - Category: content - Owner: Pulse (/pulse) When the user has an idea but isn't sure what format to use, this skill recommends a primary format plus 1-2 alternatives, anchored to the goal (reach / engagement / authority / lead-gen). Returns reasoning so the user can override. **When to use this** - user has an idea but is unsure of format - user mentions 'should this be a post or a thread', 'what format works best', 'help me pick' - user is comparing carousel vs single post vs video for the same idea **When NOT to use this** - user already knows the format → skip to the targeted skill - user wants the actual content written → use content-pulse **Example prompts** - `should this be a carousel or a video` - `what format for my launch announcement` - `help me pick: thread or single post` - `format recommendation for engagement` **Inputs** - **idea** — the content idea / angle - **goal** — reach, engagement, authority, lead-gen - **platform** — optional — linkedin, twitter, instagram, etc. **Output:** Recommended format + 1-2 alternatives + rationale per option. --- ## Blog Outline > Structured blog outline — H2/H3 hierarchy, key points per section, SEO-aware. - URL: https://app.51ultron.com/docs/skills/blog-outline - Kind: Skill - Category: content - Owner: Pulse (/pulse) Produces a long-form content outline: title options, intro framing, H2/H3 hierarchy with key points per section, SEO keyword integration, suggested word count per section, conclusion + CTA. Ready to hand to a writer or expand into the full draft. **When to use this** - user wants a blog post outline / article structure - user mentions 'blog post', 'article', 'content brief', 'SEO content', 'long-form' - user wants the structure before writing the full piece - user wants an SEO-aware brief **When NOT to use this** - user wants short-form content (LinkedIn / Twitter / IG) → use linkedin-post / social-content - user wants the full blog written, not outlined → expand outline OR use content-pulse - user wants visual / canvas output → use canvas-intelligence **Example prompts** - `blog outline on cold email best practices` - `SEO content brief for 'how to find leads'` - `long-form article structure` - `blog brief for our launch announcement` **Inputs** - **topic** — the blog subject - **keywords** — optional SEO target keywords - **length** — optional target word count - **audience** — optional reader persona **Output:** Title options + intro framing + H2/H3 structure + key points + SEO + suggested word counts. --- ## Canvas Intelligence > Build visual deliverables — carousels, charts, dashboards, comparison tables, SWOT, funnels, timelines. - URL: https://app.51ultron.com/docs/skills/canvas-intelligence - Kind: Skill - Category: design - Owner: Pulse (/pulse) Generates structured visual output for the canvas system. Picks the right canvas type (carousel / chart / dashboard / comparison / SWOT / funnel / timeline / infographic), gathers data, and produces the canvas-tagged JSON the chat layer auto-builds. Used when the user wants 'show me' / 'visualize' / 'create a canvas'. **When to use this** - user wants a visual deliverable - user mentions 'show me', 'visualize', 'create a canvas', 'build a chart', 'make an infographic', 'carousel' - user wants a comparison table, SWOT, funnel, timeline, dashboard - user wants to turn data into a visual artifact **When NOT to use this** - user wants text-only content → use content-pulse / linkedin-post / etc. - user wants a Striker call deliverable (lead magnet / blueprint) → use striker-pre-call / striker-post-call **Example prompts** - `make me a carousel about cold email mistakes` - `build a comparison chart for these tools` - `SWOT for our product` - `funnel canvas for our outbound process` - `timeline of our funding rounds` - `infographic on AI agent adoption` **Inputs** - **visual_type** — optional explicit type — carousel, chart, SWOT, etc. - **topic** — what the visual is about - **data** — optional explicit data to visualize **Output:** Canvas-tagged JSON the chat layer auto-builds into a rendered visual. --- ## Pipeline Review > Sales pipeline summary — deal health, forecast, stage breakdown, stuck-deal flags. - URL: https://app.51ultron.com/docs/skills/pipeline-review - Kind: Skill - Category: ops - Owner: Shared Services Reads the user's deals/leads/CRM state, summarizes by stage, flags stuck deals, calculates basic forecast metrics (close rate, deal velocity, weighted pipeline), and recommends actions for the riskiest deals. **When to use this** - user wants a pipeline summary or deal-health analysis - user mentions 'pipeline', 'deals', 'forecast', 'close rate', 'deal velocity', 'stuck deals' - user wants to know which deals need attention - user is doing weekly pipeline review **When NOT to use this** - user wants a daily summary across all domains → use morning-briefing - user wants to score leads (not deals) → use lead-scoring - user wants to research a specific deal's prospect → use company-deep-dive **Example prompts** - `pipeline review` - `deal health check` - `what's stuck in my pipeline` - `forecast for the quarter` - `weekly pipeline summary` **Inputs** - **window** — optional time window (default current quarter) - **filter** — optional segment / owner / stage filter **Output:** Stage-by-stage summary + forecast + stuck-deal list + recommended actions per risky deal. --- ## Morning Briefing > Daily standup — what needs attention today across leads, tasks, signals, and the calendar. - URL: https://app.51ultron.com/docs/skills/morning-briefing - Kind: Skill - Category: ops - Owner: Shared Services The user's daily 'start my day' digest. Pulls today's tasks, overdue items, fresh leads, hot signals, and upcoming meetings into one short briefing. Cheap (Haiku-class) so it can be run automatically every morning. **When to use this** - user wants a daily briefing or morning summary - user mentions 'morning briefing', 'daily update', 'what's happening today', 'priorities', 'start my day' - user wants to know what to focus on today - user wants the daily standup digest **When NOT to use this** - user wants pipeline-only review → use pipeline-review - user wants a long research piece → use company-deep-dive - user wants market trends digest → use trend-feed **Example prompts** - `morning briefing` - `what's on my plate today` - `start my day` - `daily summary` - `what should I focus on this morning` **Inputs** - **include_calendar** — optional, default true if calendar is connected **Output:** Compact briefing: today's tasks, overdue items, fresh leads, hot signals, upcoming meetings. --- ## Contract Draft > Draft a new NDA, MSA, SOW, consulting, or freelance contract — signature-ready markdown. - URL: https://app.51ultron.com/docs/skills/contract-draft - Kind: Skill - Category: legal - Owner: Counsel (/legal) Produces a complete, signature-ready contract draft. Picks template by type (NDA / MSA / SOW / consulting / freelance), fills in user + counterparty details, applies sensible defaults for term / payment / IP / liability, and outputs markdown the user can route through `crm_manage_agreements` for PDF + e-signature. **When to use this** - user wants to DRAFT a new contract from scratch - user mentions 'draft an NDA', 'write a contract', 'I need an MSA', 'generate a service agreement' - user is sending a contract to a counterparty - user wants a template-based contract for a specific type **When NOT to use this** - user has a contract from someone else and wants it reviewed → use contract-review - user wants to compare two versions → use contract-compare - user needs counter-proposal language for a flagged clause → use contract-negotiate - user wants a single-clause variant → use clause-library **Example prompts** - `draft an NDA for Acme` - `MSA for our consulting engagement with Stripe` - `SOW for the Q1 project` - `freelance contract template` - `generate a consulting agreement` **Inputs** - **contract_type** — NDA, MSA, SOW, consulting, freelance - **counterparty** — the other party (name + entity type) - **key_terms** — optional overrides for term / payment / scope / IP / liability **Output:** Signature-ready markdown contract + a CTA to route through crm_manage_agreements for PDF + e-sig. --- ## Contract Review > Review an inbound contract for risks, missing protections, and prioritized changes. Returns Safety Score (A+/F). - URL: https://app.51ultron.com/docs/skills/contract-review - Kind: Skill - Category: legal - Owner: Counsel (/legal) Weighted multi-dimension review (lifted from ai-legal-claude pattern). Each pass scores a dimension (liability / IP / termination / payment / IP residuals), produces a clause-by-clause analysis, and rolls up to a Contract Safety Score (A+/F). The passes run as normal sequential turns. **When to use this** - user uploads or pastes a contract and wants it reviewed - user mentions 'review this contract', 'is this NDA fair', 'should I sign this', 'red-line this' - user wants a Contract Safety Score before signing - user wants risks flagged with priority **When NOT to use this** - user wants to draft a NEW contract → use contract-draft - user wants to compare V1 vs V2 → use contract-compare - user wants counter-proposal language for ONE flagged clause → use contract-negotiate - user wants a plain-English summary, not risk review → use contract-plain **Example prompts** - `review this contract` - `is this NDA fair` - `should I sign this MSA` - `red-line this agreement` - `Contract Safety Score for this SOW` **Inputs** - **contract_text** — the contract to review (text or URL) - **perspective** — optional — buyer, seller, or 'mutual' **Output:** Contract Safety Score (A+/F) + clause-by-clause risk analysis + prioritized list of changes to request. --- ## Clause Library > Standard / aggressive / defensive variants of a specific clause — indemnity, liability cap, IP, etc. - URL: https://app.51ultron.com/docs/skills/clause-library - Kind: Skill - Category: legal - Owner: Counsel (/legal) When the user wants a specific clause type, returns 2-3 production-ready variants tagged by stance: standard / aggressive (favorable to the user) / defensive (favorable to the counterparty). Covers indemnity, liability cap, termination, IP assignment, non-compete, dispute resolution, etc. **When to use this** - user wants a specific clause type with variants - user mentions 'how should the indemnity clause read', 'give me a stronger termination clause' - user is filling a gap in a draft and wants pre-made language - user wants standard vs aggressive vs defensive for one clause **When NOT to use this** - user wants the WHOLE contract → use contract-draft - user wants to push back on a clause from the other side → use contract-negotiate - user wants a plain-English explanation → use contract-plain **Example prompts** - `give me an indemnity clause` - `stronger termination clause` - `non-compete language` - `liability cap variants` - `IP assignment clause for a freelancer` **Inputs** - **clause_type** — indemnity, liability cap, termination, IP, non-compete, etc. - **stance** — optional — standard, aggressive, defensive, or 'all three' - **context** — optional contract type for context **Output:** 2-3 clause variants per request, tagged by stance, with rationale. --- ## Contract Negotiate > Counter-proposals for a flagged clause — opening / landing / fallback + cover note for opposing counsel. - URL: https://app.51ultron.com/docs/skills/contract-negotiate - Kind: Skill - Category: legal - Owner: Counsel (/legal) When the user has a flagged clause (from review or counterparty redline), produces 3 counter-proposal options: opening (best case), landing (likely outcome), fallback (worst acceptable). Plus a cover note ready to email back to opposing counsel. **When to use this** - user has a flagged clause and needs counter-proposal language - user mentions 'how should I push back on this', 'rewrite this to be more favorable', 'send me a counter' - user is in a redline back-and-forth - user wants tiered options (best / likely / fallback) **When NOT to use this** - user wants to review the contract first → use contract-review - user wants generic clause variants → use clause-library - user wants to draft a NEW contract → use contract-draft **Example prompts** - `they pushed back on the IP clause — counter` - `how should I rewrite the indemnity to be more favorable` - `send me a counter-proposal for this redline` - `negotiate the liability cap` **Inputs** - **clause** — the flagged clause language - **objection** — what's wrong with it from the user's perspective - **stance** — optional — buyer, seller, mutual **Output:** 3 counter-proposals (opening / landing / fallback) + a cover note for the email back to opposing counsel. --- ## Contract Compare > Diff two versions of a contract — Material / Minor / Cosmetic categorization. - URL: https://app.51ultron.com/docs/skills/contract-compare - Kind: Skill - Category: legal - Owner: Counsel (/legal) Takes two contract versions and produces a structured diff. Each change tagged Material (significant — needs review), Minor (small — should still flag), or Cosmetic (formatting / typos). Useful for fast triage when the other side returns a redline. **When to use this** - user has two versions of a contract and wants the diff - user mentions 'compare these two versions', 'show me their redlines', 'diff these contracts' - user wants to triage a returned redline quickly **When NOT to use this** - user has ONE contract and wants it reviewed → use contract-review - user wants counter-proposal language → use contract-negotiate - user wants a plain-English summary → use contract-plain **Example prompts** - `compare V1 and V2 of this MSA` - `diff these two NDAs` - `show me what changed in their redline` - `version compare` **Inputs** - **version_a** — the original contract text - **version_b** — the modified contract text **Output:** Structured diff with each change tagged Material / Minor / Cosmetic + rationale per Material change. --- ## Contract Plain-English > Translate legalese to plain English — what each side commits to and where to push back. - URL: https://app.51ultron.com/docs/skills/contract-plain - Kind: Skill - Category: legal - Owner: Counsel (/legal) Plain-English walkthrough of a contract or specific clause. Focuses on: what each side actually commits to, what's worth pushing back on, common gotchas. Not a risk-scored review — just clarity. **When to use this** - user wants a plain-English summary of a contract or clause - user mentions 'translate this contract from legalese', 'TL;DR this MSA', 'explain this clause' - user is non-lawyer and wants the contract gist before signing **When NOT to use this** - user wants a full risk review with Safety Score → use contract-review - user wants to compare two versions → use contract-compare - user wants negotiation counter-proposals → use contract-negotiate **Example prompts** - `TL;DR this MSA` - `translate this NDA from legalese` - `what does this clause actually mean` - `plain-English version of this contract` **Inputs** - **contract_text** — the contract or clause to translate - **depth** — optional — full contract walkthrough or single-clause focus **Output:** Plain-English walkthrough — commitments per side, things to negotiate, common gotchas. --- ## Ads — ABM Sync > Closed-loop ABM — coordinate ad engagement signals back into outbound sales. - URL: https://app.51ultron.com/docs/skills/ads-abm-sync - Kind: Skill - Category: paid-ads - Owner: Amplify (/amplify) Plans the closed-loop: target account list → LinkedIn ads → engagement-tier scoring (T1 / T2 / T3) → triggered outbound from Specter sequences. Returns a tier-based signal-to-action map that the user can run with their existing sequences. **When to use this** - user wants to coordinate ads with outbound sales (ABM) - user mentions 'sync ads with outbound', 'engagement-triggered outreach', 'who from my target list engaged' - user has Specter sequences AND a target account list - user wants account-based marketing closed-loop **When NOT to use this** - user has no outbound sequences yet → run /specter sequences first - user wants only ad audiences (not coordination) → use ads-audiences - user wants outbound only (not ads) → use specter-cold-outreach **Example prompts** - `sync my ads with outbound` - `ABM closed-loop plan` - `engagement-triggered outreach from LinkedIn ads` - `who from my target list engaged` **Inputs** - **target_accounts** — list of target accounts already in the user's CRM - **sequence_ids** — optional Specter sequences to trigger from **Output:** Tier-based signal-to-action map (T1 hot / T2 warm / T3 awareness) with triggered-outreach plan per tier. --- ## Ads Launch Plan > The pre-flight package for a paid campaign: funnel architecture, audience layers with exclusions, and starting bids with pacing. - URL: https://app.51ultron.com/docs/skills/ads-launch-plan - Kind: Skill - Category: paid-ads - Owner: Amplify (/amplify) Designs the campaign before a dollar is spent. Plan mode produces the one-page funnel blueprint with a launch checklist; audiences mode builds up to three targeting plans with mandatory exclusions; bidding mode sets one strategy and starting bid per layer with a pacing table. **When to use this** - you are setting up a new paid campaign on LinkedIn, Meta, or Google - you have a budget and need the funnel, targeting, and bids designed - your targeting feels too broad or too narrow before launch **When NOT to use this** - a live campaign is underperforming: use ads-analytics - you need the copy or creative itself: use ads-assets **Example prompts** - `Set up a LinkedIn campaign with $3k/month` - `Who should I target for our RevOps tool?` - `Sanity-check my bids before launch` **Output:** One-page launch blueprint, audience plans with exclusions, bid and pacing tables. --- ## Ads Assets > The ads themselves: platform-budgeted copy variants and creative direction briefs, three per layer, banlist enforced. - URL: https://app.51ultron.com/docs/skills/ads-assets - Kind: Skill - Category: paid-ads - Owner: Amplify (/amplify) Produces ad copy and creative briefs. Copy mode returns up to three variants per asset per funnel layer inside platform character budgets, plus one A/B pair; brief mode returns three half-page design briefs with a refresh cadence. **When to use this** - you need ad copy written or rewritten - you want creative direction for statics, video, or carousels - your team asks how often to refresh creative **When NOT to use this** - funnel structure, targeting, or budgets: use ads-launch-plan - diagnosing why an ad underperforms: use ads-analytics **Example prompts** - `Write LinkedIn ad copy for our launch` - `Give me three creative briefs for retargeting` - `This headline is flat, fix it` **Output:** Copy variants per layer inside character budgets, or three design briefs with refresh cadence. --- ## Ads Analytics > Instrument tracking and attribution, then diagnose live campaigns with one-lever discipline and honest kill criteria. - URL: https://app.51ultron.com/docs/skills/ads-analytics - Kind: Skill - Category: paid-ads - Owner: Amplify (/amplify) Two modes. Instrument sets up conversion tracking, CAPI, and attribution so reporting tells the truth. Diagnose walks the decision tree on a live campaign, refuses to optimize on noise (under 7 days or 50 conversions), and returns the top three findings with exactly one lever to pull and a 3-7 day test plan. **When to use this** - conversion tracking or attribution needs setting up or fixing - a live campaign is underperforming and you want a diagnosis - you are deciding whether to kill a campaign **When NOT to use this** - designing a new campaign, audiences, or bids: use ads-launch-plan - producing copy or creative: use ads-assets **Example prompts** - `CPL doubled this week, what do I change?` - `Set up LinkedIn conversion tracking properly` - `Should I kill this campaign or wait?` **Output:** Instrumentation plan, or a diagnosis with one lever and a bounded test plan. --- ## Spin up a Web App > Scaffold a fresh web app in /work, start the dev server, surface the live preview URL. - URL: https://app.51ultron.com/docs/skills/spin-up-web-app - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) End-to-end web-app bootstrap. Picks the right framework (Next.js, Vite, Astro, plain HTML) based on the user's hint, scaffolds into /work, installs deps, starts the dev server bound to 0.0.0.0:3000 via start-dev-server, then calls shell_preview_info so the Computer tab auto-flips to the live iframe. Reads dev-preview/SKILL.md inside the container for routing rules. **When to use this** - user wants to start a new web app from scratch - user says 'build me a Next.js dashboard' / 'spin up a Vite app' / 'scaffold an Astro site' - user has a stack in mind and wants the live preview running fast - user wants to iterate on a UI with hot reload in the Computer tab **When NOT to use this** - user wants to clone an existing live site → use clone-website - user wants to deploy something that already exists → use deploy-to-cloudflare - user wants a non-web project (CLI, library) → use shell_exec directly **Example prompts** - `spin up a Next.js dashboard with shadcn` - `scaffold a Vite + React app for me` - `make me an Astro blog template` - `start a fresh Next.js project with Tailwind` **Inputs** - **stack** — framework + UI lib hints (e.g. 'next + shadcn') - **name** — optional /work subdirectory name **Output:** Live preview URL bound to 0.0.0.0:3000 + path inside /work where the project lives. --- ## Clone a Website > Fire ultron-capture: spider the site, deploy a branded clone, surface the URL when ready. - URL: https://app.51ultron.com/docs/skills/clone-website - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) Long-running async clone via the ultron-capture pipeline. Call the start_site_capture chat tool with the URL. It returns IMMEDIATELY with a job_id — the user sees the job in the Background Jobs panel. Simple sites finish in ~30s, full SPAs (Apple-scale) take up to 40 min. When done, bg_trigger fires and you'll resume the conversation with the branded deploy URL (.clones.51ultron.com) and the slug. From there: if the user wants to edit, shell_exec wget the deployed site into /work// and use the edit-html-clone skill for surgical edits, then ultron-deploy to ship the modified version. END THE TURN immediately after calling start_site_capture — do NOT poll, do NOT loop, do NOT use the in-container ultron-clone (that's the broken v1 that misses SPA content). **When to use this** - user wants to clone an existing website - user says 'rebrand ' / 'make a copy of stripe.com' / 'mirror linear.app' - user wants to swap copy on a competitor's landing and ship a variant - user wants the site as a starting point for edits + deploy to Pages **When NOT to use this** - user wants to scaffold a NEW web app from scratch → use spin-up-web-app - user wants to scrape data only (no rebuild/deploy) → use scrape_website - user wants only a screenshot of a page → run ultron-screenshot via shell_exec **Example prompts** - `clone https://linear.app and rebrand it for our launch` - `make a copy of stripe.com landing page` - `mirror flora.ai so I can edit the copy` - `rebrand vercel.com with our colors and ship to Pages` **Inputs** - **url** — the URL to capture - **branch** — optional kebab-case branch name for the Pages deploy **Output:** job_id (visible in Background Jobs panel). Branded URL surfaces via bg_trigger when capture completes (30s–40min depending on site size). --- ## Build an Agent > Spin up an autonomous agent on the agent-runtime satellite — try templates first, custom build only if no fit. - URL: https://app.51ultron.com/docs/skills/build-agent - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) Two-phase agent builder. Phase 1: agent templates lists 25 curated patterns (support, research, digest, review, status-page-monitor, etc.); if any match the user's goal, use it. Phase 2 (fallback): follows build-agent/SKILL.md to scaffold from scratch with the right tool palette (compose-tools/SKILL.md keeps the integration list minimum-viable to avoid model confusion). Returns the agent id and an inspect URL. **When to use this** - user wants an autonomous agent that runs on a schedule - user says 'build me a support bot' / 'I want an agent that monitors X' / 'set up an automation that runs every morning' - user has a recurring workflow to automate **When NOT to use this** - user wants a one-off task (no scheduling, no autonomy) → use whatever tool fits - user wants to inspect an EXISTING agent → run `agent show ` directly **Example prompts** - `build me a customer-support agent that watches new tickets` - `set up an agent that gives me a daily digest of competitor moves` - `I want an agent that monitors my status page and alerts on outages` - `create an agent that reviews PRs` **Inputs** - **goal** — what the agent should do, in one sentence - **schedule** — optional cron or natural-language frequency **Output:** Agent id, the chosen template (or 'custom'), and the agent-runtime inspect URL. --- ## Deploy to Cloudflare > Ship the current /work to Pages, Workers, or Containers — then verify the live URL. - URL: https://app.51ultron.com/docs/skills/deploy-to-cloudflare - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) Cloudflare-only deploy pipeline. Detects project shape (static HTML → Pages, single Worker script → wrangler deploy, Dockerfile → Containers). Consults cf-containers/SKILL.md for container-specific quirks (envVar timing, rollout caching). After wrangler succeeds, runs the verify-deploy skill: hits the live URL, captures status code, screenshots the homepage with ultron-screenshot, surfaces any runtime errors that wrangler couldn't catch. **When to use this** - user wants to deploy to Cloudflare (Pages, Workers, or Containers) - user says 'ship this' / 'deploy this' / 'put it on Cloudflare' when the project lives in /work - user already has a wrangler.toml and just wants it deployed **When NOT to use this** - user wants Vercel / Fly / Render → run their CLI via shell_exec directly - user wants to clone-and-deploy a site → use clone-website (which includes ultron-deploy) - user wants to deploy from a GitHub repo via CI → use push-to-github + wire CF Pages to it **Example prompts** - `deploy this to Cloudflare Pages` - `ship the worker in /work` - `push this container to Cloudflare` - `deploy and verify it's live` **Inputs** - **target** — pages | workers | containers (auto-detected from project shape if omitted) - **branch** — for Pages, optional deploy branch name **Output:** Live URL + verify-deploy report (status code, screenshot path, any runtime errors). --- ## Push to GitHub > Initialize /work as a git repo, create the GitHub repo, push, return the URL. - URL: https://app.51ultron.com/docs/skills/push-to-github - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) Wraps ultron-github-push: inits /work as a git repo if needed, commits the current state, creates the GitHub repo under the user's account using GH_TOKEN, pushes the initial commit, returns the repo URL. Optionally opens a draft PR if --pr is requested. **When to use this** - user wants to push /work to a new GitHub repo - user says 'put this on GitHub' / 'create a repo for this' / 'open-source this' - user wants the code shareable / under version control externally **When NOT to use this** - user already has a remote and just wants to commit + push → use shell_exec with git directly - user wants to fork an existing repo → use shell_exec with gh CLI directly **Example prompts** - `push this to GitHub as a new repo named cool-app` - `create a GitHub repo for this and push it` - `open-source this with the name foo-bar` **Inputs** - **repo_name** — kebab-case GitHub repo name - **private** — true|false (default false) **Output:** GitHub repo URL. --- ## Swap Page Sections > Reorder
blocks in a minified HTML file with byte-level safety. - URL: https://app.51ultron.com/docs/skills/swap-page-sections - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) Uses the swap-sections tool (and the matching SKILL.md) to swap two
blocks in /work/index.html. Handles the minified-into-one-line case via grep anchors + byte offsets so the swap doesn't corrupt the document. **When to use this** - user wants to reorder sections in a cloned landing page - user says 'move the pricing section above the features' / 'swap section A and B' **When NOT to use this** - user wants to edit text inside a section → use shell_write / string-editing skill instead - the HTML is not minified to one line → use a normal editor; this is overkill **Example prompts** - `swap the hero and the testimonials sections` - `move the pricing block above the features block in index.html` **Inputs** - **section_a** — anchor or index of first section - **section_b** — anchor or index of second **Output:** Updated /work/index.html with the two sections swapped; diff summary. --- ## Refactor Across Files > Multi-file string operations with quote and escape safety — uses bun -e, not sed. - URL: https://app.51ultron.com/docs/skills/refactor-across-files - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) Follows the string-editing/SKILL.md guidance: when editing literal strings that contain quotes, backslashes, or special characters across many files, sed corrupts the output. This capability uses bun -e (or shell_write per file) so the strings land verbatim. Best for renames, import-path migrations, API-key swaps, mass copy edits. **When to use this** - user wants to rename a symbol across 10+ files - user says 'replace every X with Y in /work' / 'change all imports from a to b' - user is doing a mass edit involving quotes or template strings **When NOT to use this** - user wants ONE file edited surgically → use shell_write or minimal-patch - user wants to fix a single bug → use minimal-patch instead **Example prompts** - `rename foo to bar across the whole codebase` - `swap every @company/old-pkg import for @company/new-pkg` - `change all API endpoints from /v1/ to /v2/` **Inputs** - **target** — the symbol/string to find - **replacement** — what to swap it with - **scope** — glob pattern (default **/*) **Output:** Diff summary per file + total file count. --- ## Iterate Until Green > Pin a goal + verifiable check; loop attempt → test → diagnose → fix until the check exits 0. - URL: https://app.51ultron.com/docs/skills/iterate-until-green - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) Implements goal-driven-loop/SKILL.md. The user supplies a goal and a verifiable shell check (a one-liner that exits 0 on success, non-zero with diagnostic output on failure). The model runs an outer loop: attempt → run check → if fail, read failure output, propose surgical patch, apply, retry. Capped at 10 iterations to avoid runaway. Best for build/test/typecheck driven work. **When to use this** - user wants something to 'just work' and has a clear pass/fail criterion - user says 'make tests pass' / 'get this building' / 'fix until typecheck is green' - open-ended task where the only spec is a runnable verification **When NOT to use this** - no verifiable check exists → ask the user for one first - task is a single bug fix → use minimal-patch directly **Example prompts** - `make the tests pass` - `iterate until pnpm typecheck is green` - `get the build working — npm run build should exit 0` **Inputs** - **goal** — what success looks like in one sentence - **check_command** — shell one-liner that exits 0 on success **Output:** Iterations taken, final check output, list of files touched. --- ## Debug a CF Worker > Tail logs, decode CF exception codes (1101 uncaught, 1102 CPU), trace which boundary is failing. - URL: https://app.51ultron.com/docs/skills/debug-cf-worker - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) Implements debug-cf-worker/SKILL.md. Tails the live Worker logs via wrangler tail, captures the exception, decodes the error code, identifies which boundary is at fault (Worker code, DO RPC, container fetch, fetch upstream). When the failure spans multiple layers, layers in instrument-then-fix/SKILL.md: insert raw passthrough diagnostics at each boundary to pinpoint the broken layer before patching. **When to use this** - user reports a Worker returning 500 - user says 'CF exception 1101' or 'error code 1102' - Worker works locally but fails when deployed - DO RPC failing or container fetch returning HTML error page **When NOT to use this** - Worker hasn't been deployed yet → use deploy-to-cloudflare and verify-deploy first - user wants to debug a non-Worker process (Next.js dev server, etc.) → use shell_exec + iterate-until-green **Example prompts** - `this Worker is returning 500 — debug it` - `Worker exception 1101 in production` - `the DO is throwing 'Cannot read property of undefined' under load` **Inputs** - **worker_name** — wrangler.toml `name` field - **error_signal** — URL that triggers it, or exception code **Output:** Root-cause analysis + the patch applied + post-fix verify result. --- ## Provision a Database > Spin up D1, R2, KV, Vectorize, Hyperdrive, Supabase, or Neon — and wire the binding/connection string. - URL: https://app.51ultron.com/docs/skills/provision-database - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) Wraps provision-database/SKILL.md + the provision-db tool. Picks the right database for the access pattern (D1 for SQL on CF, KV for hot key/value, Vectorize for embeddings, R2 for blobs, Hyperdrive for external Postgres pooling, Supabase for auth + Postgres + Storage, Neon for serverless Postgres). Provisions via wrangler or the provider's API and prints the binding stanza (for CF) or connection string (for external) ready to paste into wrangler.toml or .env. **When to use this** - user wants a database for a new app - user says 'I need a DB' / 'add auth/users' / 'store X persistently' - user wants to add embeddings/vector search to an app **When NOT to use this** - user already has a DB and wants to query it → use shell_exec with the right CLI - user wants to migrate between providers → that's a separate workflow, not provisioning **Example prompts** - `provision a D1 database called users-db` - `I need vector search — set up Vectorize` - `spin up Supabase for auth and Postgres` - `give me an R2 bucket for uploads` **Inputs** - **type** — d1 | r2 | kv | vectorize | hyperdrive | supabase | neon - **name** — resource name **Output:** Binding stanza (wrangler.toml format) or connection string (for external DBs). --- ## Verify a Deployment > Hit the live URL, capture status code, screenshot the homepage — prove runtime works (deploy success is not). - URL: https://app.51ultron.com/docs/skills/verify-deployment - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) Implements verify-deploy/SKILL.md. After any deploy (CF, Vercel, Fly, anything), runs curl against the live URL to capture status code + headers, hits critical endpoints listed by the user, screenshots the homepage with ultron-screenshot. Surfaces runtime errors that wrangler/vercel-cli can't catch (broken bindings, missing env, 500 from app code). Returns a structured pass/fail report. **When to use this** - right after any deploy - user says 'is it live?' / 'check the deploy' / 'verify production' - deploy succeeded but the user wants proof the app actually works **When NOT to use this** - before any deploy has happened → deploy first - the deploy itself failed → debug the deploy CLI output instead **Example prompts** - `verify my-app.pages.dev is live and serving` - `check the deploy at https://foo.workers.dev — should return 200` - `is the production deploy actually working?` **Inputs** - **url** — the live URL to probe - **endpoints** — optional list of critical paths to also hit **Output:** Status codes per endpoint + screenshot path + pass/fail summary. --- ## Schedule a Workflow > Build a CF Workflow — multi-day durable orchestration with sleeps + retries that survive redeploys. - URL: https://app.51ultron.com/docs/skills/schedule-workflow - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) Implements workflows/SKILL.md. Scaffolds a Cloudflare Workflow that runs on a schedule, sleeps arbitrary durations between steps (hours, days, weeks), retries failed steps with backoff, and resumes after Worker redeploys. Each step must be idempotent (re-running it produces the same result). Best for slow pipelines, daily digests, multi-day onboarding sequences, retry-tomorrow-on-failure jobs. **When to use this** - user wants something to run daily / weekly / on a cadence - user says 'run this every morning' / 'schedule this workflow' / 'do X then sleep a day then Y' - pipeline needs to retry on failure with backoff **When NOT to use this** - user wants a one-shot task → just run it - the schedule is intra-second / sub-minute → use a Durable Object alarm or queue instead - user wants an autonomous agent on the agent-runtime → use build-agent **Example prompts** - `schedule a daily digest of competitor news at 9am UTC` - `build a workflow that emails the user every Monday with last week's metrics` - `run scraper, sleep 1 day, summarize, sleep 1 day, send report` **Inputs** - **purpose** — what the workflow does in one sentence - **cadence** — how often it runs (e.g. '@daily 9am UTC') **Output:** Workflow file path in /work + deploy instructions + the cron trigger config. --- ## Website Audit > Lighthouse-style audit: perf, SEO, security, accessibility — driven by Playwright in the container. - URL: https://app.51ultron.com/docs/skills/website-audit - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) Runs ultron-render with Playwright against the target URL, capturing performance metrics (LCP, CLS, TTFB), scanning meta tags for SEO issues, checking common OWASP-Top-10 surface (CSP, HSTS, XSS, mixed content), and auditing accessibility (alt text, heading hierarchy, color contrast). Outputs a structured report with severity-ranked findings. **When to use this** - user wants to audit their site or a competitor's - user says 'audit this site' / 'how does X score on Lighthouse?' / 'check this site for security issues' - user is shipping and wants pre-launch quality check **When NOT to use this** - user only wants a screenshot → use ultron-screenshot directly - user wants to fix issues, not just identify them → audit first, then minimal-patch **Example prompts** - `audit stripe.com — perf, SEO, security` - `how does my landing score on Lighthouse?` - `run a full accessibility audit on this site` **Inputs** - **url** — site to audit - **focus** — optional comma-list: perf,seo,security,a11y (default: all) **Output:** Severity-ranked report (P0/P1/P2/P3) with metric values + concrete remediation per finding. --- ## TDD a Feature > Red → green → refactor. Failing test first, minimum-code to pass, refactor only after green. - URL: https://app.51ultron.com/docs/skills/tdd-feature - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) Strict TDD discipline. Writes the failing test via shell_write, runs it via shell_exec, watches it fail (red), implements the minimum code to make it pass (green), only then considers refactoring (per minimal-patch/SKILL.md). Forbidden to skip the red phase. Forbidden to refactor while red. **When to use this** - user wants to build a new feature TDD-style - user says 'TDD this' / 'red green refactor' / 'write the test first' - user wants strong test coverage from day one **When NOT to use this** - feature has no testable interface → use spin-up-web-app or shell_exec directly - user is fixing a bug, not adding a feature → use minimal-patch + add regression test **Example prompts** - `TDD a function that validates email addresses` - `build the rate-limiter TDD style` - `red-green-refactor a new endpoint /api/foo` **Inputs** - **feature** — what to build in one sentence - **test_runner** — pnpm test | bun test | etc. **Output:** Implementation file + test file + the green test-runner output. --- ## Security Review > OWASP Top 10 walk against the codebase with concrete proof-of-concept exploits per finding. - URL: https://app.51ultron.com/docs/skills/security-review - Kind: Shell tool - Category: engineering - Owner: Sentinel (/sentinel) Walks OWASP Top 10 against the codebase in /work via shell_read + grep + targeted endpoint probes. For every finding, builds a CONCRETE proof-of-concept (the curl command, the payload, the exact line in the code) — vague 'might be vulnerable to XSS' is forbidden. Proposes the minimum patch per finding (minimal-patch discipline). **When to use this** - user wants a security review of their code - user says 'security audit' / 'check this for vulnerabilities' / 'OWASP review' - user is shipping to production and wants a pre-launch security check **When NOT to use this** - user wants a LIVE site audited (not source) → use website-audit with focus=security - user wants to fix one specific vuln → use minimal-patch directly **Example prompts** - `security review of /work — OWASP Top 10` - `audit my API for SQL injection` - `check this auth flow for vulnerabilities` **Inputs** - **scope** — path or glob to audit (default /work) **Output:** Per-finding report: file:line, severity, PoC, proposed minimal patch. --- # Free resources library 163 free guides, playbooks, and deployable agent blueprints. Browse them all at https://app.51ultron.com/resources. - [5 Ultron skills every beginner should install](https://app.51ultron.com/resources/5-claude-skills-every-beginner-should-install) — The right five skills quietly turn Ultron from a smart chat into a machine that builds real files, plans real projects, and ships real work. - [Build an Ultron skill that analyzes prediction markets (educational build guide)](https://app.51ultron.com/resources/build-a-claude-skill-that-analyzes-prediction-markets-educational-build-guide) — Ultron can read live prediction markets and turn crowd-money odds into real analysis, and this guide shows you exactly how to build that skill. - [The Ultron starter pack: your first five minutes](https://app.51ultron.com/resources/claude-starter-pack) — The founders who win with AI did not find a secret prompt, they set Ultron up once and reached for it every single day. - [The AI System Starter Pack: build your first AI system with Ultron](https://app.51ultron.com/resources/ai-system-starter-pack) — An AI system is not a pile of tools or a folder of saved prompts, it is a small set of files that make Ultron work like your team, and you can stand - [How to master Ultron in one weekend](https://app.51ultron.com/resources/master-claude-weekend) — Most people open Ultron in a browser tab, ask a few questions, decide AI is not ready yet, and never learn they were using the weakest version of it - [Set up Ultron in one weekend: the C.A.S.T method](https://app.51ultron.com/resources/cast-method) — If AI has ever felt average to you, it was almost never the model, it was that you asked a stranger to read your mind. - [Email Agent](https://app.51ultron.com/resources/email-agent) — Read, classify, and auto-reply to incoming emails with AI-driven classification and outreach drafting. - [Lead Qualifier](https://app.51ultron.com/resources/lead-qualifier) — Research prospects by email or name, qualify them against custom criteria, and push alerts to Slack. - [Data Extractor](https://app.51ultron.com/resources/data-extractor) — Pull structured JSON data from any website or SPA using headless browser automation. - [Support Bot](https://app.51ultron.com/resources/support-bot) — Answer customer questions from your docs with automatic email escalation for unanswered queries. - [Landing Writer](https://app.51ultron.com/resources/landing-writer) — Generate conversion-optimized landing page copy in three distinct voices with live preview. - [Invoice Parser](https://app.51ultron.com/resources/invoice-parser) — Extract fields from invoices, match against PO database, and push to accounting with human approval. - [Resume Screener](https://app.51ultron.com/resources/resume-screener) — Score and rank candidates against any job description with tiered evaluation and structured feedback. - [Code Reviewer](https://app.51ultron.com/resources/code-reviewer) — Review pull requests for bugs, security issues, and style with severity-ranked inline comments. - [SQL Assistant](https://app.51ultron.com/resources/sql-assistant) — Query and edit databases using natural language with a Supabase-style table editor interface. - [Thread Writer](https://app.51ultron.com/resources/thread-writer) — Turn ideas into viral X/Twitter threads with voice matching, code blocks, and media attachments. - [Form Filler](https://app.51ultron.com/resources/form-filler) — Auto-complete web forms from context using headless browser automation and structured extraction. - [Uptime Monitor](https://app.51ultron.com/resources/uptime-monitor) — Monitor websites and APIs with intelligent diff detection, scheduled checks, and alert routing. - [Meeting Notes](https://app.51ultron.com/resources/meeting-notes) — Save, search, update, and organize notes with a persistent notebook assistant powered by Convex. - [Animation Builder](https://app.51ultron.com/resources/animation-builder) — Generate Lottie/Bodymovin animations from plain-language descriptions with live preview. - [Video Editor](https://app.51ultron.com/resources/video-editor) — Edit video clips with AI instructions using a non-linear timeline editor powered by ffmpeg. - [AEO Tracker](https://app.51ultron.com/resources/aeo-tracker) — Track how often your brand surfaces in answers from Perplexity, Gemini, ChatGPT, and Claude with mention rate, position, sentiment, and competitor data. - [ICP Architect](https://app.51ultron.com/resources/icp-architect) — Build a deep ideal customer profile from product description and market signals with firmographics, pain points, buying triggers, and TAM estimate. - [Cold Outreach Writer](https://app.51ultron.com/resources/cold-outreach-writer) — Research a prospect and draft three cold-email variants with subject lines scored by predicted reply rate before you send. - [Competitor Ad Intel](https://app.51ultron.com/resources/competitor-ad-intel) — Scrape competitor ad creative from Meta Ads Library and Google Ads Transparency, then break down hooks, offers, and visual patterns into a strategy doc. - [SEO Content Engine](https://app.51ultron.com/resources/seo-content-engine) — Turn a target keyword into a full content brief with SERP analysis, intent, outline, internal links, and meta ready to hand to a writer. - [Buyer Persona Generator](https://app.51ultron.com/resources/buyer-persona-generator) — Synthesize qualitative interviews, sales-call notes, and product analytics into rich buyer personas with quotes, objections, and channel preferences. - [Trigger Event Hunter](https://app.51ultron.com/resources/trigger-event-hunter) — Monitor target accounts for funding rounds, executive hires, product launches, and news mentions, then push qualified buying signals to Slack. - [Topical Authority Mapper](https://app.51ultron.com/resources/topical-authority-mapper) — Build a content cluster map with pillar pages, supporting articles, and an internal-link graph to dominate a niche on AI search and Google. - [Deep Research Agent](https://app.51ultron.com/resources/deep-research-agent) — Multi-source research dossier on any company, market, or person, pulling from web, news, filings, and forums into a fact-checked memo. - [Pricing Researcher](https://app.51ultron.com/resources/pricing-researcher) — Scrape competitor pricing pages, decode tier structure and gating, and recommend a positioning move backed by real comparison data. - [Sales Battlecard Generator](https://app.51ultron.com/resources/sales-battlecard-generator) — Generate per-competitor battlecards with objection scripts, win themes, and trap-setter questions reps can use on live calls. - [Customer Health Monitor](https://app.51ultron.com/resources/customer-health-monitor) — Score customer accounts on churn risk using product usage, support tickets, and engagement signals, then alert CSMs before renewals slip. - [Content Repurposer](https://app.51ultron.com/resources/content-repurposer) — Turn one long-form input into blog, Twitter, LinkedIn, and email variants matched to each platform's voice and length. - [CRM Hygiene Bot](https://app.51ultron.com/resources/crm-hygiene-bot) — Dedupe leads, enrich missing firmographics, flag stale records, and queue cleanup tasks so your pipeline reports stay trustworthy. - [Backlink Hunter](https://app.51ultron.com/resources/backlink-hunter) — Find link-building opportunities by topic, qualify each prospect by domain authority and relevance, and draft personalized outreach. - [LinkedIn Engager](https://app.51ultron.com/resources/linkedin-engager) — Find relevant posts in your niche, draft contextual comments that add real value, and surface high-engagement threads worth joining. - [Investor Update Drafter](https://app.51ultron.com/resources/investor-update-drafter) — Compose a monthly investor update from metrics, milestones, and qualitative notes in your voice, ready to send via email. - [Onboarding Sequencer](https://app.51ultron.com/resources/onboarding-sequencer) — Send behavior-triggered onboarding emails that adapt to user actions, hand off cold accounts to sales, and stop spamming activated ones. - [Review Aggregator](https://app.51ultron.com/resources/review-aggregator) — Pull customer reviews from G2, Capterra, and Trustpilot, cluster by theme, and surface the strongest pull-quotes for marketing use. - [Customer Interview Synthesizer](https://app.51ultron.com/resources/customer-interview-synthesizer) — Take a stack of raw interview transcripts and synthesize jobs-to-be-done, recurring objections, and willingness-to-pay signals. - [Webinar Brief Builder](https://app.51ultron.com/resources/webinar-brief-builder) — Generate a webinar title, hook, outline, slides plan, and 4-week promotion calendar from a topic and target audience. - [Podcast Show Notes](https://app.51ultron.com/resources/podcast-show-notes) — Turn a podcast transcript into structured show notes with chapter timestamps, key quotes, guest bio, and shareable takeaways. - [Voice Receptionist](https://app.51ultron.com/resources/voice-receptionist) — Answer every inbound call, book appointments, and take messages around the clock so a local business never sends a caller to voicemail. - [Website Concierge](https://app.51ultron.com/resources/website-concierge) — An embedded chat widget that answers visitor questions, captures leads, and books appointments straight from a local business website. - [Build a bedtime story app in your own voice with Ultron](https://app.51ultron.com/resources/bedtime-story-app-your-own-voice) — Your own voice can read the bedtime story on the nights you cannot be there, and Ultron builds the whole app end to end. - [Set up your first Ultron agent (human-in-the-loop starter guide)](https://app.51ultron.com/resources/set-up-your-first-claude-agent-human-in-the-loop-starter-guide) — An agent is just Ultron with one job, and this weekend you can set up your first one and grow into running a whole team of them. - [The $10,000 website: build yours end-to-end in Ultron](https://app.51ultron.com/resources/10k-website-emergent) — With Ultron you can build a $10,000-quality website end-to-end: it writes the plan, builds the actual site, puts it live on the internet, and hands - [Build a $5k scroll animation with zero code, end to end in Ultron](https://app.51ultron.com/resources/5k-scroll-animation) — You can build the $5,000 scroll animation from Apple's product pages yourself, end to end in Ultron, without writing a single line of code. - [Build a $10k-looking dashboard in Ultron, no code](https://app.51ultron.com/resources/antigravity-dashboard) — You can build a $10k-looking, interactive dashboard end to end in Ultron with zero code, just by describing the job and refining it live. - [Make Ultron build a site that doesn't look AI: the 4-agent team](https://app.51ultron.com/resources/site-not-look-ai) — Make Ultron build a custom site that actually converts by running it as a four-agent team, instead of letting it spit out another purple, AI-looking - [Build your own AI agent, safely, in Ultron](https://app.51ultron.com/resources/build-ai-agent-safely) — You can build a real AI agent that books tables, fills forms, and runs weekly research for you, and you can do it safely inside Ultron without ever - [Rebuild any website you love in Ultron](https://app.51ultron.com/resources/rebuild-website-claude) — Hand Ultron a screenshot and the copied styles from any website you love, and it will rebuild the look and self-correct until the site is yours. - [Build a SaaS app with Ultron (full guide)](https://app.51ultron.com/resources/saas-app-blink) — You can turn an idea into a real, working SaaS app this week, entirely inside Ultron, even if you have never written a line of code. - [Build a mobile app with Ultron (full guide)](https://app.51ultron.com/resources/mobile-app-blink) — You can build a real phone app for iPhone and Android this week, entirely inside Ultron, with no Mac, no code, and no developer. - [How to replace a $120k dev with Claude Code](https://app.51ultron.com/resources/claude-code-playbook) — The complete playbook for running Claude Code as 10 persistent AI departments that ship production code, manage repos, and handle DevOps without supervision. - [The prompt standard that separates amateurs from pros](https://app.51ultron.com/resources/prompt-engineering-standard) — Stop writing prompts that return generic slop. Learn the 7-layer prompt architecture used by production AI systems to guarantee consistent, high-fidelity outputs every single time. - [How to orchestrate Claude Code subagent swarms](https://app.51ultron.com/resources/claude-code-subagents) — The architecture for splitting complex coding tasks into parallel subagent workflows where a lead agent delegates, a research agent investigates, and build agents ship code simultaneously. - [How to use Claude Code for free in production](https://app.51ultron.com/resources/free-claude-code) — The complete setup guide for running Claude Code without a paid subscription by leveraging API credits, free tier allocations, and open source tooling that replicates the full coding agent experience. - [The 8 Claude Code templates that ship features in minutes](https://app.51ultron.com/resources/claude-code-templates) — Pre-built, copy-paste-ready Claude Code configurations for authentication flows, Stripe billing, CRUD APIs, database migrations, testing suites, CI/CD pipelines, landing pages, and admin dashboards. - [How to unlock Claude superpowers with MCP tools](https://app.51ultron.com/resources/claude-101-superpowers-mcp) — The practical guide to Model Context Protocol that teaches you how to give Claude access to your file system, databases, APIs, and browser so it stops being a chatbot and starts being an autonomous operator. - [How to turn AI skills into a plugin marketplace](https://app.51ultron.com/resources/agent-factory-plugins) — The architecture for packaging your best AI workflows as installable plugins that other businesses can buy, deploy, and run inside their own agent infrastructure. - [How to build AI agents that improve themselves](https://app.51ultron.com/resources/agent-self-evolution) — The self-evolution architecture where agents analyze their own outputs, identify failure patterns, update their system prompts, and get measurably better over time without human intervention. - [The lightweight Claude Code setup for any machine](https://app.51ultron.com/resources/claude-code-light-list) — A minimal terminal-based Claude Code configuration that gives you the full autonomous coding experience on low-spec machines, remote servers, and headless environments without VS Code or Cursor. - [How to give AI agents persistent memory across tasks](https://app.51ultron.com/resources/agent-knowledge-layer) — The knowledge layer architecture that stores agent learnings, context, and preferences in a structured database so every agent remembers past interactions and gets smarter with every task. - [How to build agents that evolve their own strategy](https://app.51ultron.com/resources/evonexus-orchestration) — The evolutionary orchestration engine where agent teams test different coordination strategies, measure outcomes, and automatically adopt the patterns that produce the best results for your business. - [How to build a specialist agent hive that wins 6x](https://app.51ultron.com/resources/multi-agent-hive) — The hive architecture where 5 to 8 narrow-scope agents collaborate on complex tasks, each contributing domain expertise that a single generalist agent cannot match. - [Bulk ad voiceovers: let Ultron write ten and voice them](https://app.51ultron.com/resources/bulk-ad-voiceovers-claude-fish-audio) — The fastest way to find a winning ad is to test ten at once, not agonize over one. - [Generate 50 UGC video ads overnight with Ultron](https://app.51ultron.com/resources/generate-50-ugc-ads-overnight-claude-higgsfield) — You can test more ad ideas in a single night with Ultron than most teams test in a month. - [Run a marketer-grade Instagram growth audit with Ultron](https://app.51ultron.com/resources/claude-instagram-growth-audit-prompt-kit) — Your Instagram is not underperforming because of the algorithm, it is underperforming because you are measuring the wrong things. - [Point Ultron at your Meta ads and let it run the reporting](https://app.51ultron.com/resources/connect-claude-to-meta-ads-mcp-setup-starter-prompts) — The reporting and diagnosis grind on your Meta ads is exactly the work you should never touch again. - [5 ways Ultron replaces your Meta ads agency](https://app.51ultron.com/resources/meta-ads-skills) — Your Meta ads agency bills you five to ten thousand dollars a month for work Ultron finishes before lunch. - [Connect your Meta ads to Ultron and stop living in Ads Manager](https://app.51ultron.com/resources/connect-meta-ads) — Ultron connects straight to your Meta ad account, pulls the reports, launches the campaigns, and moves the budgets, so you stop living inside Ads - [5 jobs Ultron runs to replace your Meta ads agency](https://app.51ultron.com/resources/claude-skills-meta-ads) — Your Meta ads agency charges 5,000 to 10,000 dollars a month for work Ultron runs in minutes. - [How Ultron Took an Instagram Account From 0 to 30k Followers in 3 Weeks](https://app.51ultron.com/resources/0-to-30k-ig) — The Instagram algorithm does not owe you reach, you earn it by stacking the right systems on top of the right rules, and this is the exact 4-system - [The Ultron Meta ads playbook: 3 systems from $1M+ in ad spend](https://app.51ultron.com/resources/claude-meta-ads-playbook) — Most Meta ad accounts waste 30 to 50 percent of their budget on ads that should already be off, and Ultron can find every one of them in about thirty - [Build your entire lead funnel from one request with Ultron](https://app.51ultron.com/resources/lead-funnel-perspective) — Your ad budget is not leaking on the ad. - [The 8 Ultron agents that run a marketing system: the full build guide](https://app.51ultron.com/resources/8-agents-marketing) — Most founders use AI for one task at a time. - [How to publish 30 SEO articles a month with zero writers](https://app.51ultron.com/resources/ai-content-engine) — The content engine that turns a keyword list into fully researched, brand-voiced, SEO-optimized articles at $5 per piece instead of $350 from a freelancer. - [How to book 40 B2B meetings a month on autopilot](https://app.51ultron.com/resources/b2b-local-outreach) — The pipeline that scrapes every business in your target radius, visits their website, scores buying intent, and sends emails that look completely handwritten. - [How to dominate Google and AI search engines](https://app.51ultron.com/resources/geo-seo-claude-optimization) — The dual optimization framework that ranks your content on traditional Google while simultaneously training AI models like ChatGPT and Perplexity to cite your brand as the authoritative source. - [How to find the angle that makes your product sell](https://app.51ultron.com/resources/positioning-angles) — The systematic framework for discovering the one strategic angle that transforms your product from 'nice to have' to 'obvious choice' in your target customer's mind within 30 days. - [The 47 best places to launch your product in 2026](https://app.51ultron.com/resources/places-to-launch) — A ranked database of every launch platform, community, and directory that drives real signups, sorted by traffic quality, competition level, and expected conversion rate for B2B and B2C products. - [How to produce 90 days of content in one afternoon](https://app.51ultron.com/resources/content-system) — The full production pipeline that takes a single long-form piece and atomizes it into 90 social posts, 12 email newsletters, 6 blog articles, and 30 short-form video scripts without repeating a single idea. - [The content engine that runs on autopilot: Ultron is the brain](https://app.51ultron.com/resources/content-engine) — Most creators drown in tools when what they really need is one brain that plans the work and then does it. - [The faceless YouTube system: let Ultron pick the niche and write the scripts](https://app.51ultron.com/resources/faceless-youtube-system-claude-fish-audio) — Faceless YouTube channels do not fail because of the missing face, they fail because they pick the wrong niche and sound like a robot. - [How to make the text wrap effect](https://app.51ultron.com/resources/text-wrap-effect) — Two Ultron prompts let you build the text-wrap effect and the binary rain version yourself, with no coding needed. - [The Instagram carousel generator: the Ultron project prompt](https://app.51ultron.com/resources/carousel-generator) — One sentence in, a fully designed, swipeable, export-ready Instagram carousel out, without ever opening a design tool. - [A month of on-brand carousels, designed by Ultron](https://app.51ultron.com/resources/month-of-carousels) — A single carousel should take you minutes, not a whole afternoon. - [How to grow your business alone like a 20-person team](https://app.51ultron.com/resources/grow-like-a-20-person-team) — Most founders are still hiring for roles AI replaced 12 months ago. Discover the $200 stack that replaces a $30k/month team. - [How to run profitable ads without an agency or media buyer](https://app.51ultron.com/resources/autonomous-ai-ad-engine) — The $350/mo system that generates 50 ad variations daily, kills underperformers automatically, and outspends agencies charging $5k/mo retainers. - [How to stress test any business decision with AI](https://app.51ultron.com/resources/boardroom-executive-simulator) — Simulate a full board meeting with an AI CFO, CMO, and market analyst that debate your strategy, attack your blind spots, and surface risks you missed. - [How to ship a product in 48 hours from an idea](https://app.51ultron.com/resources/ceos-ai-operating-system) — The operating system that captures your raw voice memos, synthesizes them into PRDs and task lists, and scaffolds the code before you finish your morning coffee. - [How to find hidden revenue in your Shopify data](https://app.51ultron.com/resources/claude-ecom-business-intelligence) — Pipe your raw order data into Claude and get instant answers about which products, bundles, and discount codes are secretly destroying your margins. - [How to launch a product in 3 days flat](https://app.51ultron.com/resources/claude-gtm-bible) — Feed Claude one context document about your product and watch it generate the landing page, email sequences, ad scripts, and sales playbook in a single afternoon. - [How to steal your competitors' best Meta ads](https://app.51ultron.com/resources/meta-ads-spy-system) — The intelligence system that monitors competitor ad libraries daily, extracts their winning hooks and creatives, and feeds the insights directly into your own campaign engine. - [The 12 AI skills every founder needs in 2026](https://app.51ultron.com/resources/founder-playbook-ai-skills) — A structured skill tree that takes you from basic prompting to deploying autonomous agent swarms, with specific tools and practice exercises for each level. - [How to detect trends 6 weeks before competitors](https://app.51ultron.com/resources/signal-hunter-trend-scouting) — The AI intelligence system that monitors Reddit, Twitter, patent filings, and VC deal flow in real time to surface weak signals that become major trends before mainstream awareness. - [How to run an AI company from a single laptop](https://app.51ultron.com/resources/portable-agent-company) — The architecture for a location-independent business where every department is an AI agent, every process is automated, and the entire operation fits in a backpack with zero office infrastructure. - [The complete AI system from zero to full automation](https://app.51ultron.com/resources/complete-system) — The end-to-end blueprint that connects every tool, agent, and workflow into a single unified system. From lead generation to delivery to billing, every step is automated and every agent knows its role. - [How to stress test your pitch before meeting investors](https://app.51ultron.com/resources/vc-investment-simulation) — Run your pitch through 5 AI investor personas that ask the exact questions real VCs ask, score your deck on 12 dimensions, and generate a pre-mortem analysis of every way the deal could fall apart. - [The 9 decision charts that fix your GTM strategy](https://app.51ultron.com/resources/gtm-decisions-charts) — Pre-built decision trees that tell you exactly which GTM motion to run based on your ACV, market size, competition density, and current traction. No more copying strategies that do not fit your business. - [How to deploy AI agents with enterprise compliance](https://app.51ultron.com/resources/enterprise-governance-harness) — The enterprise-grade governance harness that satisfies SOC 2, GDPR, and HIPAA requirements while still letting AI agents operate autonomously within defined compliance boundaries. - [How to launch a startup in 30 days with AI agents](https://app.51ultron.com/resources/startupkit-orchestrator) — The day-by-day launch blueprint that uses AI agents for market validation, MVP development, landing page creation, content production, and initial customer acquisition across a compressed 30-day timeline. - [How to find $10k in hidden automation savings](https://app.51ultron.com/resources/ai-audit-process) — The step-by-step framework that maps every manual workflow in your company and calculates the exact ROI of replacing each one with an AI agent. - [How to review contracts in seconds for free](https://app.51ultron.com/resources/ai-legal-assistant) — Train an AI legal assistant on your company risk profile that reads 50-page NDAs in 12 seconds, flags dangerous clauses, and generates plain-English summaries. - [How to merge 15 tools into one autonomous hub](https://app.51ultron.com/resources/business-os-automation-hub) — The architecture that unifies your CRM, billing, project management, and communication into a single data layer that AI agents can actually read and act on. - [How to build a company OS in a single weekend](https://app.51ultron.com/resources/company-os-starter-kit) — The Notion-based OS template that structures your entire business into machine-readable databases so AI agents can navigate clients, projects, and finances autonomously. - [The 14-tool stack that replaces a 10-person team](https://app.51ultron.com/resources/solo-toolkit-founder-stack) — A curated, tested stack of 14 tools that covers sales, marketing, engineering, finance, legal, support, and operations for solo founders running lean AI-native businesses. - [How to run your entire business from one dashboard](https://app.51ultron.com/resources/founder-os-build-system) — A unified command center that consolidates revenue metrics, pipeline status, content performance, agent activity, and financial health into a single real-time dashboard you check for 10 minutes each morning. - [How to govern 20+ AI agents without losing control](https://app.51ultron.com/resources/agentic-network-governance) — The governance framework for multi-agent systems that defines permission hierarchies, communication protocols, conflict resolution, and audit trails so your agent network scales without breaking. - [How to architect a Claude-powered business OS](https://app.51ultron.com/resources/claude-ops-business-os) — The operational architecture that uses Claude as the intelligence layer across every business function, with structured handoffs, escalation rules, and human-in-the-loop checkpoints at critical decisions. - [How to control your AI agents without micromanaging](https://app.51ultron.com/resources/human-control-plane) — The human control plane architecture that gives you strategic oversight of your entire agent network through exception-based alerts, approval queues, and kill switches without requiring you to monitor every task. - [Make videos from code with Ultron: the full setup](https://app.51ultron.com/resources/edit-videos-with-claude-code-and-remotion-the-full-setup) — Ultron turns a plain-English prompt into finished motion graphics, building animated charts, kinetic text, and bulk ad variants straight from code. - [10 work processes to hand to Ultron and standardize across your team](https://app.51ultron.com/resources/10-ready-to-use-claude-skills-for-work) — The processes your team repeats every week are the ones you should never do by hand again. - [Build a focus system with Ultron (anti-scroll prompts)](https://app.51ultron.com/resources/build-a-focus-system-with-claude-anti-scroll-prompts) — Willpower will never beat an infinite feed, so stop fighting it and have Ultron build a focus system that physically gets in the way. - [10 things to build with Ultron](https://app.51ultron.com/resources/build-with-artifacts) — Ultron does not just answer you, it opens a workspace beside your chat and builds a real, editable thing that you shape by talking. - [The study stack: 3 Ultron prompts that make you smarter (not lazier)](https://app.51ultron.com/resources/study-stack) — These three Ultron prompts force you to actually learn, instead of the passive ask-AI-paste-answer loop that quietly makes you a worse thinker. - [The truth prompt: how to make Ultron stop making things up](https://app.51ultron.com/resources/truth-prompt) — The truth prompt is a custom instruction that forces Ultron to calibrate its confidence and admit when it does not know, instead of confidently - [6 ways to make Ultron pressure-test your ideas instead of flattering you](https://app.51ultron.com/resources/stop-claude-yes-man) — Most AI tells you your idea is brilliant right up until it fails. - [The Ultron Council: run any big decision past five specialist agents](https://app.51ultron.com/resources/claude-council) — One confident AI answer is exactly how smart founders talk themselves into expensive mistakes. - [Make any hard decision in 5 minutes: the 10/10/10 coach](https://app.51ultron.com/resources/10-10-10-coach) — Ultron can become a 10/10/10 decision coach that makes any hard decision in five minutes and won't let you hide from it. - [Win any argument: the steel-man generator](https://app.51ultron.com/resources/steel-man-generator) — Ultron can become a steel-man generator that builds the other side stronger than the people who actually hold it, then dares you to refute it. - [5 Ultron prompts that make you sharper, not lazier](https://app.51ultron.com/resources/sharper-not-lazier) — These 5 Ultron prompts turn it into a sparring partner instead of an answer machine, so you get sharper while it helps, not lazier. - [How Ultron Actually Shrinks Your To-Do List in 30 Seconds](https://app.51ultron.com/resources/30-second-todo) — Your to-do list never shrinks because everything on it feels important, and the one job you cannot do for yourself is decide what to cut. - [5 ways Ultron pressure-tests your decisions instead of agreeing with you](https://app.51ultron.com/resources/brutally-honest-decisions) — Most AI tells you your plan is brilliant right up until reality proves otherwise. - [Stop Ultron being a yes-man: the devil's advocate setup](https://app.51ultron.com/resources/devils-advocate) — An AI that agrees with everything you say is quietly making your ideas worse. - [How to automate email from triage to reply](https://app.51ultron.com/resources/claude-email-system) — The system that reads every incoming email, classifies it by urgency and type, drafts contextual replies in your voice, and only pings you for decisions that actually matter. - [The 10 Claude prompts that replaced $15k a month](https://app.51ultron.com/resources/top-10-claude-prompts) — Production-tested prompts for financial modeling, legal review, market research, copywriting, data analysis, technical architecture, hiring, competitive intel, sales scripts, and investor decks. - [The CLEAR framework that kills 80% of AI hallucinations](https://app.51ultron.com/resources/clear-context-os) — A 5-layer context management system (Context, Limits, Examples, Actions, Rules) that structures every AI interaction to produce reliable, grounded outputs instead of confident-sounding hallucinations. - [How to make Claude read all your company docs](https://app.51ultron.com/resources/company-docs-mcp) — Set up a Company Docs MCP server that indexes your SOPs, wikis, handbooks, and knowledge bases so Claude can answer any employee question with accurate, sourced information from your actual documents. - [5 secret Ultron prompts](https://app.51ultron.com/resources/5-secret-claude-prompts) — These five prompts turn Ultron into a brutally honest thinking partner, ready to copy and paste straight into your next chat. - [100 secret Ultron codes (the cheat-sheet edition)](https://app.51ultron.com/resources/100-claude-codes) — These 100 Ultron codes are keywords you drop into a prompt to completely change what Ultron gives you back, the same way a cheat code flips a video - [Turn Ultron into a powerhouse: 5 marketing codes + 100 for every field](https://app.51ultron.com/resources/claude-powerhouse-codes) — The reason Ultron hands you generic copy is never the model, it is the vague brief behind it, and every code below fixes that by loading a proven - [Make Ultron work while you sleep: the overnight autopilot setup](https://app.51ultron.com/resources/claude-overnight) — Your business should keep moving after you close the laptop, and with Ultron it does. - [The perfect Ultron brief template (and 5 worked examples)](https://app.51ultron.com/resources/great-prompt-template) — The gap between a generic result and a great one is never the AI, it is the brief you gave it. - [Save any content workflow as a one-word command in Ultron](https://app.51ultron.com/resources/claude-skills-command) — Stop retyping the same brand prompt on every request. - [The Ultron skills actually worth installing](https://app.51ultron.com/resources/skills-worth-installing) — Skills are the fastest way to make Ultron dramatically better at one specific job, and these are the ones actually worth installing. - [24 things to install in Ultron: the power-user setup guide](https://app.51ultron.com/resources/24-things-install-claude) — Ultron gets dramatically more powerful the moment you start adding packs, skills, and connectors, and here are 24 worth installing. - [Design with Ultron: setup plus 3 power moves most people miss](https://app.51ultron.com/resources/claude-design-setup-plus-3-power-moves-most-people-miss) — Ultron turns a blank page into a finished, on-brand asset while you describe what you want in plain English. - [Build a professional-looking website with Ultron (no design background)](https://app.51ultron.com/resources/build-a-professional-looking-website-with-claude-code-no-design-background) — A professional-looking website is not about design talent, it is about giving Ultron the right design rules and real content, then letting it build - [7 Ultron skills for creative teams (full setup guide)](https://app.51ultron.com/resources/claude-creative-teams) — These 7 Ultron skills run an entire content engine end to end, research, writing, voice consistency, video, audio, ad intelligence, and persistent - [How to turn Ultron into your full-time designer](https://app.51ultron.com/resources/claude-full-time-designer) — Ultron builds real landing pages, decks, PDF lead magnets, and ad-style screenshots from one chat, with no designer on payroll. - [Turn any photo into a $10k-looking product ad in Ultron](https://app.51ultron.com/resources/product-ad-reve) — Most product photos look flat because you skipped the art direction, not because of your phone, and Ultron fixes that by directing the shot and - [Design posters like a world-class designer in Ultron](https://app.51ultron.com/resources/posters-reve) — The fix for generic AI posters is the trick real designers use: feed Ultron a look you already love and tell it to borrow the art direction, not copy - [Build a hand-controlled brain visual with Ultron](https://app.51ultron.com/resources/claude-touch-design) — You can turn your webcam into a glowing, hand-controlled map of your own ideas, and build the whole thing by pasting one prompt into Ultron. - [How to generate on-brand design assets in 12 seconds flat](https://app.51ultron.com/resources/ai-native-design-systems) — Convert your brand guidelines into programmatic tokens so AI can produce pixel-perfect social posts, pitch decks, and landing pages without a designer. - [4 Ultron skills that make your resume unrejectable](https://app.51ultron.com/resources/resume-unrejectable) — Four Ultron skills, run in order, turn a resume that keeps getting ignored into one that clears the applicant tracking system, matches what - [Connect Ultron to Indeed and stop scrolling job boards](https://app.51ultron.com/resources/claude-indeed) — Connect Ultron to Indeed and it pulls live listings, reads your profile, and shortlists the roles actually worth your time, so you stop scrolling job - [Run your whole job search on autopilot: 3 prompts](https://app.51ultron.com/resources/job-search-autopilot) — These 3 Ultron prompts run your whole job search as one system: a tracker that catches everything, a resume re-fit per job, and batch outreach to - [Tailor your resume to any job in 10 minutes: 3-agent prompt](https://app.51ultron.com/resources/tailor-resume) — One Ultron prompt spins up three agents that tailor your resume to any job in about ten minutes, so it surfaces past the ATS instead of getting - [Your 4-agent salary negotiation team](https://app.51ultron.com/resources/salary-negotiation-team) — One Ultron chat spins up a four-agent team that finds your real market rate, sets your numbers, scripts your exact words, and drills you on every - [Get a reply from a lead that went cold: 4 follow-ups Ultron writes for you](https://app.51ultron.com/resources/job-ghosted-followups) — The deal you think is dead is usually one good follow-up away from alive. - [How to use Ultron to beat the ATS: your 4-agent resume team](https://app.51ultron.com/resources/beat-ats-4-agent) — Your resume is screened by software before a human ever sees it, and this four-agent team you run inside Ultron ranks you high enough that a - [Beat the ATS: your 6-agent resume team in Ultron](https://app.51ultron.com/resources/beat-ats-6-agent) — Your resume is not read by a person first, it is ranked by software, and this six-agent team you run inside Ultron is built to surface you to the top - [100 careers AI won't replace, and how to pick yours with Ultron](https://app.51ultron.com/resources/100-careers-ai-wont-replace) — The jobs AI can't take all rest on six human skills, and your smartest career move is to find which one is already yours, not to out-code the machine. - [5 AI skills worth real money in 2026 (and how to learn each with Ultron)](https://app.51ultron.com/resources/5-ai-skills-money) — The people making real money with AI in 2026 did not learn everything; they picked one skill, went deep, and let one tool do the heavy lifting. - [Turn Ultron into your interior designer](https://app.51ultron.com/resources/claude-interior-designer) — Ultron can redesign any room from a single photo and tell you exactly what to buy, so you can plan the whole space before you spend a dollar. - [Travel with Ultron: your stress-free trip setup guide](https://app.51ultron.com/resources/travel-with-claude) — One Ultron setup turns every trip from a chaos of browser tabs, apps, and half-remembered bookings into a single concierge that plans your days - [Turn Ultron into a fitness coach that reads your real health data](https://app.51ultron.com/resources/claude-fitness-coach) — Ultron can become a fitness coach that reads your real health data and adapts your plan day to day around your actual numbers. - [Fridge to feast: the Ultron prompt that turns your fridge photo into 16+ meals](https://app.51ultron.com/resources/fridge-to-feast) — Snap one photo of your fridge, hand it to Ultron, and get 16+ meal ideas back, sorted by how long they take and filtered for the way you eat. - [The Ultron glow-up prompt pack: rate your face, plan your glow up, find your colors](https://app.51ultron.com/resources/glow-up-rate-face) — Hand Ultron one clear, front facing selfie and it turns it into a clean, editorial style report with honest scores and specific, achievable ways to - [Turn selfies into a pro headshot with Ultron: the prompt pack](https://app.51ultron.com/resources/pro-headshot-chatgpt) — A few phone selfies plus one key line is all Ultron needs to rebuild a professional headshot that still looks unmistakably like you, in about thirty