Skip to content

SKILL.md

Latest commit

 

History

History
53 lines (39 loc) · 1.54 KB

SKILL.md

File metadata and controls

53 lines (39 loc) · 1.54 KB
name execution-protocol
description Validates any AI agent action before execution. Use before any Bankr trade, Base transaction, contract deployment, treasury move, governance vote, or any agent action with real-world consequences. Call POST /ep/validate with a proposal. Returns proof_hash if valid. Required before capital moves. Use when agent mentions trade, execute, deploy, validate, policy, proof, verify, authorize, approve, spend, transfer, swap, bridge, vote, govern, swarm, coordinate.
metadata
author version endpoint auth docs
Achilles / Project Olympus
1.0.0
X-Agent-Key header required

Execution Protocol

Pre-execution validation layer for AI agents. Solves Black Box Risk.

POST /ep/validate

Validate any proposed action before execution.

Auth: X-Agent-Key header.

Request:

{
  "proposal_id": "uuid",
  "agent_id": "your-agent-id",
  "asset": "BNKR",
  "direction": "buy",
  "amount_usd": 25,
  "entry_price": 0.0005,
  "stop_loss": 0.00045,
  "take_profit": 0.00065,
  "leverage": 1,
  "policy_set_id": "olympus-v1",
  "chain": "base"
}

Response: { valid, risk_score, violations[], proof_hash, plan_summary, timestamp }

Only proceed if valid: true. Attach proof_hash to ledger.

POST /ep/swarm/validate

Validate action within a multi-agent swarm. Enforces swarm-level exposure limits.

GET /ep/proof/:hash

Public proof lookup. No auth. Verifiable by anyone.

GET /ep/status

Service health. No auth.