agent-navigation-sop
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes repository contents to generate navigation instructions. If a codebase contains malicious comments or instructions intended to deceive AI agents, these might be incorporated into the generated AGENTS.md file.
- Ingestion points: The skill instructions assume the agent uses codebase analysis tools to read files within the repository.
- Boundary markers: None present to distinguish between legitimate code and potentially malicious embedded instructions in the source files being analyzed.
- Capability inventory: The skill itself does not perform network operations, file writes, or command execution; it provides instructions for outputting a markdown file.
- Sanitization: No specific sanitization or filtering logic is provided to handle untrusted input from the codebase.
- [SAFE]: The skill promotes security best practices by instructing agents to identify and document Access Constraints (read-only files/folders) and strictly forbidding the execution of blocking processes or long-running development servers.
Audit Metadata