text-diff
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes the contents of arbitrary user-provided files for comparison. This creates an attack surface where malicious instructions embedded within the files being diffed could be interpreted by the agent during the process.\n
- Ingestion points: The
scripts/text_diff.pyscript reads the full content of two target files specified as arguments.\n - Boundary markers: The script does not wrap the diff output in specific delimiters or include warnings to the agent to ignore instructions within the text.\n
- Capability inventory: The skill has the ability to read and write local files (via the
--outputflag).\n - Sanitization: No sanitization or filtering of the file content is performed before generating the diff output.
Audit Metadata