agents-md-edit
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions and associated scripts are focused on optimizing instruction files. Analysis of the audit script, decision criteria, and restructured guidelines found no evidence of prompt injection, data exfiltration, or malicious persistence.
- [COMMAND_EXECUTION]: The file scripts/test_audit.py uses subprocess.run to invoke the audit.py script for testing. The execution is limited to the local Python environment using static paths and does not accept untrusted shell input.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided AGENTS.md files, creating a potential ingestion surface for malicious instructions. Ingestion points: The agent reads user-specified Markdown files as part of the audit and edit workflow (SKILL.md). Boundary markers: The skill does not mandate specific delimiters for the ingested content. Capability inventory: The agent possesses file-writing capabilities to update the target file, while the internal audit script is strictly read-only. Sanitization: No explicit content filtering is implemented for the target file content.
Audit Metadata