docs
Pass
Audited by Gen Agent Trust Hub on Mar 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to run
devtools::document()for R projects. This is a well-known and standard development command used to generate documentation files from source code comments. - [PROMPT_INJECTION]: The skill includes an 'IMPORTANT' instruction that directs the agent to exit after completing its phases. This is a functional control instruction for the agent's workflow and does not attempt to bypass safety protocols or override system instructions.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the codebase to generate documentation. Ingestion points: Project source files and
CLAUDE.md. Boundary markers: None defined. Capability inventory: Local file modification (writing documentation) and execution ofdevtools::document(). Sanitization: No explicit sanitization or filtering of the ingested code content is specified, which represents a standard surface for indirect prompt injection, though the risk remains low given the skill's specific purpose.
Audit Metadata