cluster-corpus-by-theme
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data which may contain malicious instructions.
- Ingestion points: The skill reads every post in the
corpus/published/**directory end-to-end. - Boundary markers: The instructions do not define delimiters or warnings to ignore instructions embedded within the corpus posts.
- Capability inventory: The skill itself does not specify subprocess calls, file-write operations, or network requests, but it instructs the agent to process the full text of all files in a specific directory.
- Sanitization: There is no mention of sanitizing or escaping the content of the posts before the agent processes them.
Audit Metadata