doc-quality-review

Pass

Audited by Gen Agent Trust Hub on Sep 17, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted documentation files through sub-agents to perform quality reviews, which introduces a surface for indirect prompt injection. While this is inherent to the skill's primary function, the risk is limited as the sub-agents are tasked with evaluation rather than execution.
  • Ingestion points: Documentation files identified in the scope manifest (Phase 1) are read and passed to sub-agents (Sonnet) in Phase 2.
  • Boundary markers: The prompt template for sub-agents inlines the document content without using specific boundary delimiters or 'ignore embedded instructions' warnings to isolate the untrusted data from the evaluation logic.
  • Capability inventory: The orchestrator utilizes file system tools (Read, Grep, Glob) to locate and access files, and the Agent tool to dispatch sub-agents for analysis.
  • Sanitization: The skill does not apply sanitization, filtering, or validation to the content of the documentation files before they are interpolated into the sub-agent prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 17, 2026, 10:40 AM
Security Audit — agent-trust-hub — doc-quality-review