skill-system-auditor
Pass
Audited by Gen Agent Trust Hub on May 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run repository-specific validation scripts (e.g.,
python3 scripts/validate_skills.pyand other discovery commands). While necessary for auditing, this involves executing code that resides within the repository being audited rather than the skill itself. - [PROMPT_INJECTION]: The skill has a surface for indirect prompt injection (Category 8). It is designed to ingest and analyze multiple files from a repository, such as
SKILL.md,README.md, andAGENTS.md. These files could contain malicious instructions designed to subvert the agent's behavior during the auditing process. - Ingestion points: The agent reads repo-level guidance (
AGENTS.md,README.md) and individual skill definitions (skills/*/SKILL.md) in Step 1 and Step 2. - Boundary markers: No explicit delimiters or instructions to ignore embedded prompts are provided when processing these files.
- Capability inventory: The skill is granted
Bash,Write, andEdittools, allowing it to modify the repository based on its analysis. - Sanitization: There is no mention of sanitizing or validating the content of the audited files before processing.
Audit Metadata