knowledge-curator
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests 'candidate updates' from the current session to update durable knowledge sets and ADRs.
- Ingestion points: Candidate updates are processed in step 1 of the core process in
SKILL.md. - Boundary markers: No specific delimiters or instructions to ignore embedded commands within the candidate updates are defined.
- Capability inventory: The skill's capabilities are limited to producing Markdown artifacts and reasoning; no network access, file system modifications, or shell execution capabilities are declared.
- Sanitization: The instructions do not specify a validation or sanitization step for the ingested data.
Audit Metadata