maintain-verification-skill
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill's documentation identifies material adapted from Cursor's public plugins repository (github.com/cursor/plugins).
- [INDIRECT_PROMPT_INJECTION]: The skill processes project source code and interacts with live application outputs, representing an ingestion surface for untrusted data.
- Ingestion points: The skill reads project-local source files and processes live application recipes (SKILL.md).
- Boundary markers: No explicit delimiters or instructions to ignore embedded commands are defined.
- Capability inventory: The agent can interact with application servers, UIs, and CLI runs, and modify files within a specific verification directory (SKILL.md).
- Sanitization: No sanitization or filtering of external content is described in the instructions.
Audit Metadata