copy-critique
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill accepts untrusted user input ('Existing copy text') which is processed to generate evaluations and rewrites. This represents a standard attack surface for text-processing agents, though the risk is negligible here due to the lack of access to dangerous tools.
- Ingestion points: 'Existing copy text' input in SKILL.md.
- Boundary markers: No specific delimiters or safety instructions are defined for the input text.
- Capability inventory: The skill generates YAML-formatted text output and utilizes internal file-reading for rubrics; no subprocess, network, or file-write capabilities are present in the provided instructions.
- Sanitization: No input validation or sanitization is implemented.
Audit Metadata