sdd-archive

Pass

Audited by Gen Agent Trust Hub on Jul 21, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill incorporates an attack surface for indirect prompt injection because it processes untrusted delta specifications and integrates their content into the project's 'Source of Truth' (main specs) without sanitization.
  • Ingestion points: The agent reads delta specification files from the openspec/changes/{change-name}/specs/ directory.
  • Boundary markers: The instructions do not define clear delimiters or utilize instructions to ignore embedded commands when processing or merging the delta content.
  • Capability inventory: The skill utilizes file system capabilities including reading, writing, and moving files/directories.
  • Sanitization: There is no evidence of validation or sanitization for the requirement content being merged; it uses a text-matching approach to append, replace, or delete sections based on Markdown headers.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 21, 2026, 12:23 AM
Security Audit — agent-trust-hub — sdd-archive