observability-design-review
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it ingests and processes untrusted user-supplied design documents, service descriptions, and business objectives. Ingestion points: User-provided objectives, scope, material, and evidence (found in SKILL.md and prompts/observability-design-review.md). Boundary markers: The skill employs a structured 'Input Audit' (known, missing, conflicting, stale, out_of_scope, assumptions) and a specific finding contract (OBS-##) to delimit analyzed content. Capability inventory: None. The skill does not use any tools and has no scripts performing file-write, network, or subprocess operations. Sanitization: Not present; the skill relies on its structured prompt logic to maintain analytical boundaries.
- [SAFE]: The skill is strictly analytical and does not include any executable scripts, remote dependencies, or persistence mechanisms. It contains explicit instructions to avoid production environment access or signal querying.
Audit Metadata