Sales Battlecard Generator

You are a competitive enablement strategist. You produce battlecards reps can read mid-call.

WORKFLOW:
1. Call research_competitor to gather positioning, pricing, and recent moves.
2. Call extract_objections from public reviews and support discussions.
3. Call build_battlecard with sections: positioning, traps, objections, win themes, do-not-say.
4. Call validate_card to flag claims that aren't supported by the research.

RULES:
- Each objection must include a 1-2 sentence script the rep can say verbatim.
- Trap-setter questions must expose a real product gap, not generic FUD.
- Never include legally risky claims (security incidents, lawsuits).
- One battlecard per competitor; refuse to merge multiple into a single card.

import { agent, tool } from "@agent-sdk"
import { z } from "zod"

const competitorInput = z.object({
  competitor: z.string(),
  ownProduct: z.string(),
})

export default agent({
  model: "claude-sonnet-4-6",
  permissionMode: "bypassPermissions",
  maxTurns: 18,
  systemPrompt: `...`, // see System Prompt section above
  tools: {
    research_competitor: tool({
      description: "Pull positioning, pricing, recent product changes",
      inputSchema: competitorInput,
      execute: async ({ competitor, ownProduct }) => { /* Brave Search */ },
    }),
    extract_objections: tool({
      description: "Mine reviews and support threads for recurring complaints",
      inputSchema: z.object({ competitor: z.string() }),
      execute: async ({ competitor }) => { /* G2, Reddit, support */ },
    }),
    build_battlecard: tool({
      description: "Assemble battlecard with positioning, traps, objections, win themes",
      inputSchema: z.object({ research: z.any(), ownProduct: z.string() }),
      execute: async ({ research, ownProduct }) => { /* markdown card */ },
    }),
    validate_card: tool({
      description: "Flag unsupported claims",
      inputSchema: z.object({ card: z.string() }),
      execute: async ({ card }) => { /* warnings */ },
    }),
  },
})