brida-reflex-context-retention-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 specify reading live configuration and availability files such as /reflex/llms.txt and /reflex/agent.md from Brida's hosted infrastructure. These are vendor-owned metadata resources.
- [INDIRECT_PROMPT_INJECTION]: The skill acts as a classifier for context items, creating a surface for indirect prompt injection. 1. Ingestion points: The skill processes item summaries and task descriptions within the Reflex state. 2. Boundary markers: The instructions call for deterministic host code to apply protections before processing. 3. Capability inventory: Uses MCP, SDK, or REST to provide recommendation-only classification results. 4. Sanitization: Instructions mandate normalization of context evidence and emphasize restricted authority.
Audit Metadata