frontpage
Warn
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill requires the agent to execute shell commands found in the target repository's documentation and installation scripts to verify they work. This includes installation commands and usage snippets extracted from the project being documented.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from the project's codebase, manifests, and tests, creating a surface for indirect prompt injection that could influence the agent's behavior or output.
- Ingestion points: Codebase entry points, manifest files, CI configurations, and existing documentation as described in the Workflow and Prerequisites sections of SKILL.md.
- Boundary markers: None identified; the instructions direct the agent to ground its output in the discovered repository data without specifying delimiters for untrusted content.
- Capability inventory: File system read access (spelunk), network verification of links, and shell execution for install/usage verification.
- Sanitization: None specified for the data read from the repository before processing or execution.
Audit Metadata