relevance-coarse-filter
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes 'signals' consisting of titles and excerpts from external sources. * Ingestion points: Untrusted data enters the agent context via signal titles, excerpts, and evidence URLs as described in the 'Inputs' section of SKILL.md. * Boundary markers: The instructions do not define clear delimiters or escaping mechanisms to separate the signal content from the agent's instructions. * Capability inventory: The skill's primary function is to output a JSON object indicating a classification decision; it does not exhibit file-system, network, or subprocess capabilities. * Sanitization: There is no evidence of sanitization or filtering applied to the input signal data before it is processed by the agent.
Audit Metadata