commerce-feed-intelligence
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill references and instructs the agent to utilize local Python scripts for routing and workflow management.
- Evidence: References to
../../scripts/dynamic_router.py,../../scripts/neural_router.py, and../../scripts/skill_router.pyinSKILL.mdandreferences/skill-spec.json. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data such as commerce feed fields, merchant diagnostics, and catalog data, which could potentially contain malicious instructions.
- Ingestion points: Data inputs for 'feed fields' and 'merchant diagnostics' defined in
references/decision-model.md. - Boundary markers: The skill includes an 'Evidence discipline' section in
SKILL.mdrequiring the agent to classify claims as fact, inference, or assumption. - Capability inventory: The skill uses local scripts for routing and triggers specialist agents for synthesis.
- Sanitization: The
references/output-contract.mdmandates the separation of facts, assumptions, and unknowns in the final response.
Audit Metadata