ai-forge-audit
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bundled Node.js script (scripts/check-ecosystem.cjs) to perform roster-level coherence checks. This script is restricted to read-only filesystem operations within the repository and does not utilize dangerous primitives such as network access, subprocess spawning, or file writing. The execution is part of the skill's primary function and is performed on local, audited code.
- [PROMPT_INJECTION]: The skill processes content from other artifacts in the repository, creating a surface for indirect prompt injection.
- Ingestion points: Phase 0 reads artifact content via a Node.js script, and Phase 2 involves scanning and evaluating skill/agent files throughout the repository.
- Boundary markers: The skill does not explicitly instruct the agent to ignore malicious commands embedded in the files being audited.
- Capability inventory: The skill focuses on data extraction and report rendering; it lacks destructive write or network capabilities triggered by evaluated content.
- Sanitization: The skill isolates and extracts specific structured fields (Name, Grade, Improvements) from the evaluation output rather than raw content, which limits the attack surface for report corruption.
Audit Metadata