codepatrol-sync-state
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from GitHub pull request descriptions and documentation pages, creating a surface for indirect prompt injection.
- Ingestion points: Pull request overview sections and Cosense page content (referenced in SKILL.md under '対象のページ' and 'ページに無いpull requestを探す').
- Boundary markers: The instructions do not define explicit delimiters or instructions to ignore potential commands embedded within the pull request descriptions.
- Capability inventory: The skill is capable of reading pull request metadata and writing content to documentation pages via a secondary 'cosense skill'. It is explicitly restricted from reading source code or creating/editing pull requests.
- Sanitization: There is no evidence of sanitization, escaping, or filtering of the text retrieved from the pull request descriptions before it is used to determine state changes or written to pages.
Audit Metadata