# Skill Scan by Agentropolis

POST https://seenly-skill-scan.vercel.app/api/scan with JSON {repo | files | code | manifest, context?} → verdict (allow | review | block), risk 0-100, consensus votes, findings with file:line evidence, tools + tools_sha256, provenance, Agentropolis trust cross-check. Use it before your agent installs or connects an MCP server / skill (npx, pip install, a new SKILL.md). Static and read-only: code is fetched from GitHub / npm / PyPI only and never executed; run-time behaviour is out of scope (sandbox detonation is a separate, later tier). Payment: unpaid calls return HTTP 402 with x402 v2 PAYMENT-REQUIRED; settle only on 2xx. Invalid requests → 422 before payment; fetch failures → 4xx/5xx, not charged.

Price: $0.30 USDC per call on Base mainnet. No free calls; free sample output: https://seenly-skill-scan.vercel.app/samples/scan-sample.json

POST https://seenly-skill-scan.vercel.app/api/scan
Body: {"repo":"https://github.com/owner/repo"} or {"repo":"npm:@scope/pkg@1.2.3"} or {"files":[{"path":"server.py","content":"..."}],"manifest":{"tools":[...]}}. Optional "context": what you plan to use it for.
Returns: verdict allow|review|block, risk, consensus (rules:code, rules:manifest, 2 LLM reviewers), findings with file:line evidence, tools + tools_sha256 baseline, provenance (repo/package age), agentropolis trust cross-check.
OpenAPI: https://seenly-skill-scan.vercel.app/openapi.json
Kit: https://seenly-skill-scan.vercel.app/.well-known/kits.json
Contact: hello@agentropolis.io · https://agentropolis.io
Limits: body ≤ 240 KB; repos ≤ 12 MB compressed; version 0.1.0.