task-brief-creator
Pass
Audited by Gen Agent Trust Hub on Aug 21, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local Python scripts (
scripts/validate_brief.pyandscripts/validate_briefset.py) to verify the structural integrity of the generated plans. These scripts are bundled with the skill, operate without external dependencies, and are restricted to structural analysis of the output Markdown files. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data, such as product requirement documents and planning notes. It employs a robust verification pipeline to mitigate risks, including a mandatory ambiguity gate that halts if the input is underspecified and a 'cold-pickup' sub-agent loop that confirms the generated instructions are naturally interpreted as intended without malicious drift.
Audit Metadata