generated: '2026-07-19' method: searched description: >- Agent Skills for Knostic. One is published by Knostic itself and saved verbatim; two are generated by API Evangelist against the AgentMesh OpenAPI, with every step grounded in a real operationId from openapi/knostic-agentmesh-openapi.yml. skills: - file: knostic-extension-check.md name: extension-check method: searched source: https://github.com/knostic/extension-check-skill/blob/main/skills/extension-check/SKILL.md license: MIT official: true description: >- Knostic's own published Agent Skill. Audits installed Cursor / VS Code / VS Code Insiders extensions against AgentMesh using structured `publisher:X name:Y` searches of the rendered UI, then reports MATCH / VERSION_MISMATCH / NOT_FOUND / PARTIAL with risk labels. hosts: - Claude Code (as a plugin via the repo's bundled marketplace, or as a user skill) - Cursor (native Agent Skill in .cursor/skills/) note: >- This skill deliberately instructs agents NOT to use the AgentMesh API and to parse the rendered search UI instead; Knostic notes the skill "may not be 100% consistent" and points API consumers at the AgentMesh API for direct ingestion. - file: knostic-audit-agent-supply-chain.md name: Audit the AI agent supply chain against Knostic AgentMesh method: generated api: openapi/knostic-agentmesh-openapi.yml operations: [listSkills, getSkill, listMcp, getMcp, listExtensions, getExtension] description: >- API-first counterpart to extension-check: look up skills, MCP servers and extensions in AgentMesh by search or SHA-256 hash and read the scan verdict and risk evidence. - file: knostic-scan-on-demand.md name: Trigger and poll an on-demand AgentMesh scan method: generated api: openapi/knostic-agentmesh-openapi.yml operations: [scanSkill, scanExtension, getScan, listScans] description: >- Request a fresh scan for an unscanned or stale artifact, poll it to completion, and manage the 30/day per-key scan quota including the 429-vs-503 distinction.