Skip to content

Add vigil-security-scanner skill (onchain security on Base)#480

Open
vigilcodes wants to merge 1 commit into
BankrBot:mainfrom
vigilcodes:add-vigil-security-scanner
Open

Add vigil-security-scanner skill (onchain security on Base)#480
vigilcodes wants to merge 1 commit into
BankrBot:mainfrom
vigilcodes:add-vigil-security-scanner

Conversation

@vigilcodes

Copy link
Copy Markdown

Summary

Adds vigil-security-scanner — a keyless onchain security scanner for Base, exposed as an HTTP JSON-RPC API at mcp.vigil.codes. Lets any agent check a token or wallet before it signs.

What it does

13 read-only tools:

  • vigil_safety_score — 0-100 contract rating
  • vigil_detect_honeypot — can you sell? buy/sell tax
  • vigil_liquidity_lock — LP locked / burned / unlocked / unknown
  • vigil_scan_approvals — wallet approval risk
  • vigil_check_scam, vigil_deployer_check, vigil_token_market, vigil_monitor_wallet, vigil_wallet_report, vigil_consensus (6-source verdict), and more

Why it fits

  • Read-only & autonomous-safe. No transaction is ever signed by this skill. Revoke (write) actions are intentionally excluded.
  • Fail-safe. Undetermined results return unknown, never a false safe — important for a security tool.
  • Complements existing scanners. aeon-vuln-scanner audits code/repos; VIGIL covers onchain token/wallet security on Base. Different surface, no overlap.
  • Keyless for read-only scans — no API key required to install or use.

Testing

Tested against the live endpoint (mcp.vigil.codes) via scripts/smoke_test.sh. Health check returns 13 tools; sample calls return real verdicts (e.g. USDC → safe 92/100).

Structure

vigil-security-scanner/
├── SKILL.md
├── references/  (api-reference.md, contracts.md)
└── scripts/     (smoke_test.sh)

Homepage: https://github.com/vigilcodes/vigil-mcp

Onchain security scanner for Base — 13 read-only tools over a keyless
JSON-RPC endpoint (mcp.vigil.codes). Scan token approvals, detect
honeypots, check liquidity locks, score contracts 0-100, and get a
6-source consensus verdict.

Differentiators vs existing security skills:
- 6-source consensus with false-positive guard (single source caps at medium)
- liquidity lock detection that fails safe to 'unknown', never fake 'safe'
- keyless + free core pre-trade checks, built for autonomous agents
- MCP-native (Claude, Cursor, Aeon, any MCP client)

Includes references/api-reference.md (per-tool schema) and
scripts/smoke_test.sh (live endpoint smoke test). Read-only only —
revocation is deliberately excluded.

Tested: endpoint live (13 tools), examples verified against production.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant