ecs-security
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill utilizes a discovery-driven framework requiring the agent to process external data (user answers to eight discovery questions) to generate a customized security hardening stack. This architectural pattern creates an ingestion surface for untrusted data that influences agent output.
- Ingestion points: Discovery questions in
SKILL.md(Launch type, Compliance regime, Workload sensitivity, Audit timeline, Network topology, Secrets posture, Team skill, and Tooling baseline). - Boundary markers: The instructions do not define explicit delimiters or "ignore embedded instructions" warnings for the user-provided data.
- Capability inventory: The skill generates complex, multi-layered security recommendations and hardening roadmaps based on the provided inputs.
- Sanitization: There are no documented mechanisms for sanitizing or validating user-provided discovery answers before they are interpolated into the agent's reasoning process.
Audit Metadata