curate-carbon-component
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute project-local validation and build scripts using the npm package manager. These commands are used to check the status, build, and type-check the curated metadata.
- Evidence:
npm run skills:status,npm run build:skills,npm run test:skills,npm run build:skills -- --check, andnpm run type-check. - [INDIRECT_PROMPT_INJECTION]: The skill processes external data including component source code, Storybook stories, and MDX documentation, which represents an attack surface for indirect instructions.
- Ingestion points: Component source code files, Storybook MDX and story files, and
docs/component-metadata/README.md. - Boundary markers: The instructions do not define specific delimiters or security warnings to treat the content as data only.
- Capability inventory: The skill has access to shell execution via
npm runand read access to the local filesystem. - Sanitization: No sanitization or filtering of the ingested documentation content is specified.
Audit Metadata