cursor-maintain-verification-skill
Pass
Audited by Gen Agent Trust Hub on Oct 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill reads and processes project-local source code and feature maps to identify 'drift'. Malicious content within the project files being analyzed could potentially influence the agent's behavior during the 'Source wave' and 'Live pass' phases.
- Ingestion points: Processes project-local source files, feature files (
features/), and READMEs within the target verification skill directory (.agents/skills/verify-*/). - Boundary markers: No specific delimiters or 'ignore embedded instructions' warnings are defined for the content being processed.
- Capability inventory: Performs file-write operations (restricted to the verification skill's directory) and executes commands via CLI and local harness scripts.
- Sanitization: No explicit sanitization or validation of the ingested external content is mentioned.
- [COMMAND_EXECUTION]: The skill instructions direct the agent to execute various commands to verify application behavior.
- Evidence: Mentions running a 'real disposable CLI', performing 'syntax/version/health Doctor' checks, and executing 'harness scripts' owned by the verification skill.
Audit Metadata