ai-writing-tells
Pass
Audited by Gen Agent Trust Hub on Oct 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is a specialized utility for auditing prose quality and identifying 'AI tells'. It uses a local bash script to perform regex-based pattern matching on user-provided text files.\n- [INDIRECT_PROMPT_INJECTION]: Analysis of the attack surface for processing untrusted data.\n
- Ingestion points:
scripts/check.shreads and processes the content of files provided as arguments.\n - Boundary markers: None present; the agent is instructed to run the check on specific files.\n
- Capability inventory: The script is restricted to read-only operations using
grep,sed, andwc. It lacks network access, file-writing permissions, or the ability to execute code found within the processed data.\n - Sanitization: The content is treated as literal text for pattern matching, preventing accidental execution of embedded instructions.
Audit Metadata