tidy-code
Pass
Audited by Gen Agent Trust Hub on May 26, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted source code as its primary input, creating a surface for indirect prompt injection where malicious instructions could be embedded in code comments or string literals.\n
- Ingestion points: Source files discovered by
scripts/scan-source-files.shand accessed via theReadtool inSKILL.md.\n - Boundary markers: The orchestration logic in
SKILL.mddoes not specify delimiters or explicit instructions for sub-agents to ignore directives embedded within the analyzed code.\n - Capability inventory: The agent can read local files, write reports, and orchestrate sub-agents through the
Tasktools.\n - Sanitization: No sanitization or filtering is applied to the code content before it is passed to the analysis models.\n- [EXTERNAL_DOWNLOADS]: The
tests/run-tests.shscript downloads thebats-coretesting framework from its official repository on GitHub.
Audit Metadata