marginalia
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to use
ripgrep(rg) for code exploration and to execute local validation commands (such as type-checking or linting) defined in the project's ownpackage.jsonor configuration files. These operations are restricted to the local development environment. - [INDIRECT_PROMPT_INJECTION]: The skill analyzes local source code to generate documentation, which is a standard ingestion surface for coding tools.
- Ingestion points: The agent reads implementations, usage paths, test files, and project configuration manifests (
package.json,tsconfig.json,tsconfig*) as part of its 'Package API Workflow'. - Boundary markers: The instructions do not define specific delimiters or 'ignore' instructions for the data ingested from the codebase.
- Capability inventory: The skill allows the agent to modify local source files (inserting JSDoc) and execute local shell commands for searching and build validation.
- Sanitization: The skill enforces strict stylistic 'Comment Rules' and 'Brevity rules' to ensure generated content remains focused on technical metadata rather than executing instructions found within the code.
Audit Metadata