craque-neto
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is limited to persona-based text generation for providing project feedback. It does not include executable code, scripts, or instructions to perform network operations or access sensitive system files.
- [PROMPT_INJECTION]: The instructions contain a strong defensive posture against prompt injection. It explicitly commands the agent to treat user-provided project content as data and to ignore any embedded directives like 'ignore previous instructions' or requests to provide unearned praise.
- [INDIRECT_PROMPT_INJECTION]: The skill defines an attack surface by ingesting external project data (text, files, links). However, the risk is addressed through clear boundary markers in the instructions and the fact that the skill lacks tools or capabilities (like file writing or network sending) that could be exploited by malicious payloads in the processed data.
- [COMMAND_EXECUTION]: No shell commands, subprocess calls, or persistent mechanisms are present in the skill instructions or supporting documentation.
Audit Metadata