engineer-copilot-agent
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from the repository, which is a potential vector for indirect prompt injection.
- Ingestion points: Files located in
.github/agents/andskills/are read by the analysis scripts (discover_runtime_surface.py,validate_agent.py,analyze_agent_body.py) to extract metadata and body content. - Boundary markers: The skill instructions do not specify explicit delimiters or "ignore embedded instructions" warnings when the content of these files is presented back to the agent for improvement.
- Capability inventory: The skill possesses file-reading capabilities across the repository through its Python scripts but lacks network access (no
requestsorcurlcalls) and does not perform file-writing operations. - Sanitization: The scripts use
yaml.safe_loadfor parsing frontmatter metadata, which protects against arbitrary code execution through YAML tags, though the natural language body content is processed via regex and presented directly.
Audit Metadata