brida-reflex-test-relevance-gate
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes natural language data including change summaries, component descriptions, and test purposes to classify relevance. This creates an indirect prompt injection surface where adversarial content in software descriptions could attempt to influence the agent's decision-making logic.
- Ingestion points: Processes
changeSummary,testName, andtestPurposefields as defined inreferences/custom-reflex.jsonand described inSKILL.md. - Boundary markers: The
references/playbook.mdprovides explicit delimiters for the authority boundary, stating that the result is recommendation-only and must not waive required repository policies. - Capability inventory: The skill produces semantic recommendations used by a host runner to schedule tests; it does not have direct execution privileges for destructive actions or deployments.
- Sanitization: The instructions mandate that parsing, policy enforcement, and side effects must remain in deterministic host code, limiting the impact of any potential semantic injection.
Audit Metadata