technical-documentation
Pass
Audited by Gen Agent Trust Hub on Jun 14, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted documentation and agent instruction files (e.g., AGENTS.md, CONTRIBUTING.md, .cursorrules). These files could contain malicious instructions designed to hijack the agent's behavior during the documentation review or build process.
- Ingestion points: The skill reads various markdown and documentation framework source files from the repository.
- Boundary markers: There are no explicit instructions for the agent to use delimiters to isolate untrusted content from the system prompt.
- Capability inventory: The skill uses file-system tools (Read, Glob, Grep, LS) and has the capability to create files and symlinks in its 'apply-fixes' mode.
- Sanitization: The instructions do not specify sanitization or validation of the content within processed files before it is evaluated by the agent.
- [COMMAND_EXECUTION]: The skill executes shell-based discovery and management commands.
- The agent is instructed to use ripgrep (rg) to discover instruction files across the repository.
- The skill specifies logic to automatically create or fix symlinks (e.g., .cursor to .agents) to ensure compatibility between different AI development tools.
Audit Metadata