tk-skill-diagnose
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted data from external sources, including CI logs, PR reviews, web content, and issue trackers. It includes a specific "Retrieved Evidence Boundary" directive that mandates all such data be treated as evidence only, preventing it from overriding the agent's instructions or security boundaries.
- [COMMAND_EXECUTION]: The skill utilizes Git commands (
git ls-files,git check-ignore) to perform safety checks. These commands are used to verify that the.tigerkit/directory is properly ignored by the repository's version control before any diagnostic artifacts are written, preventing accidental leakage of transient data. - [DATA_EXFILTRATION]: While the skill manages sensitive user inputs (e.g., tokens or credentials) via temporary JSON files, it implements strict security protocols. Secret inputs are required to be stored in a dedicated directory with restricted OS permissions (0700 for directories, 0600 for files) and are explicitly excluded from being echoed in chat, logs, or command history.
Audit Metadata