Agent Workbench Pack

Raven is a signed evidence-and-receipt layer for AI agents before they touch a Solana token. This pack makes Raven usable from YOUR workbench — Codex, Claude, Cursor, Gemini, any MCP client or framework — with no human browsing required. Copy the pieces below into your agent.

Do not ask Raven whether a token is safe. Ask Raven what on-chain evidence it found, what it could not check, and whether the signed receipt is enough for your agent's policy.

1. Load context

context pack — ingest first

https://raven-launch-console.vercel.app/raven-agent-context.md

operating skill — paste into your workbench

https://raven-launch-console.vercel.app/raven-agent-skill.md

prompt recipes — 12 repeatable prompts (JSON)

https://raven-launch-console.vercel.app/prompt-recipes.json

2. Call Raven

mcp — local, free

npx -y raven-verify-mcp

hosted api — keyed

curl -s -X POST https://raven-hosted-verifier.onrender.com/verify \
  -H "x-api-key: $RAVEN_KEY" -H "content-type: application/json" \
  -d '{"mintAddress":"...","tokenProgramAddress":"..."}'

pubkey — no auth

curl -s https://raven-hosted-verifier.onrender.com/pubkey

No key? Hire RAVEN agent-to-agent: ACP offering solana_verify_token, 0.1 USDC/job, escrowed. Or request a key. Contract: /openapi.json.

3. Verify & store the receipt

receipt verification

1. GET /pubkey -> match receipt keyId (today rvk_c2997e90215279c2, ed25519)
2. Recompute replayHash (canonical JSON; recipe in repo)
3. Verify signature over utf8("raven-official-attestation")||officialAttestationHash
4. Store the FULL receipt with the action it gates (key policy: /key-policy.json)

3b. Store per the memory policy

Storage envelope schema: /receipt-schema.json · wire receipt schema (validates the receipt object itself): /receipt-wire-schema.json · policy: /receipt-memory-policy.md · adapters (all optional): /receipt-storage-adapters.json. Check stored receipts before re-calling; verify before use; storage failure never alters a verdict.

4. Decide: proceed / block / escalate

agent handoff policy

risk        -> block, or human review
warning     -> human review unless policy explicitly allows
unknowable  -> do not proceed
pass-grade + coverageGaps -> "not enough evidence for a full pass";
               proceed only if policy accepts those exact gaps
never call a token safe while gaps remain
store every receipt; escalate verification MISMATCH as an incident

False confidence is the failure mode. coverageGaps is normative — anything listed was NOT evaluated. Test us first: /evals.

Request a key Demo kit Workbench