brida-reflex-retrieval-relevance-gate

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructions in SKILL.md require the agent to fetch live configuration and availability data from external documentation paths, specifically /reflex/llms.txt, /reflex/agent.md, and /reflex/agent.json. These resources originate from the vendor's own infrastructure to ensure current feature support.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to process untrusted external data (retrieved candidate passages) and compare them against user queries.
  • Ingestion points: External text candidates are ingested via the state object defined in references/custom-reflex.json and processed at runtime.
  • Boundary markers: The skill does not explicitly define prompt boundary markers or "ignore" instructions for the ingested text.
  • Capability inventory: The skill's capabilities are restricted to semantic relevance judgment; it does not perform file writes, network exfiltration of sensitive data, or shell command execution.
  • Sanitization: No specific sanitization or filtering logic is documented for the ingested candidates.
  • The risk is minimal as the output is restricted to a binary recommendation (include/discard) within a controlled workflow.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 08:21 PM
Security Audit — agent-trust-hub — brida-reflex-retrieval-relevance-gate