yetong-title
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes a local utility
scripts/check_profile.pyto verify the structure and activation status of profile documents (person-dna.md). The script uses regex-based validation to ensure the input file matches the expected schema and contains required sections before allowing the agent to proceed. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied drafts and questions which could theoretically contain instructions intended to influence the agent's output.
- Ingestion points: User-provided text segments (existing titles, content drafts, and member questions) are used as the basis for analysis in
SKILL.md. - Boundary markers: There are no specific instructions to treat user data as untrusted text using delimiters or explicit 'ignore' warnings.
- Capability inventory: The skill uses a local validation script and generates text outputs; it does not have network access or administrative privileges.
- Sanitization: The skill does not perform explicit sanitization of the input text but relies on structural extraction (e.g., extracting a single sentence '内容承诺').
Audit Metadata