cleanup-all
Pass
Audited by Gen Agent Trust Hub on Jun 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill orchestrates the execution of eight sequential cleanup sub-tasks ('cleanup-unused', 'cleanup-cycles', etc.). It manages the control flow by invoking these local skills and halting the pipeline upon any verification failure.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) as it ingests data from the target repository to generate metrics and reports.
- Ingestion points: The skill reads file metadata (LOC, file counts) and processes output from analysis tools like 'knip' and 'madge' from the local project environment.
- Boundary markers: No explicit delimiters or instructions to the model to ignore embedded instructions within the analyzed codebase are provided.
- Capability inventory: The skill can write files (master reports in '.claude/cleanup-reports/'), execute git commands through its child workflows, and trigger subsequent automated cleanup phases.
- Sanitization: No escaping, filtering, or validation is performed on the ingested codebase metrics before they are aggregated into the markdown report.
Audit Metadata