feature-delivery

Pass

Audited by Gen Agent Trust Hub on May 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is a legitimate workflow coordinator. It utilizes the GitHub CLI (gh) for managing issues and pull requests, which is standard for development tasks. Access to external data is limited to the design documents specified by the user.
  • [PROMPT_INJECTION]: The skill processes external design documents provided via a file path. This represents an indirect prompt injection surface as the agent parses content from these documents to determine sub-tasks. However, the skill includes manual checkpoints and status verification steps (e.g., confirming the 'Revised' status) that provide oversight of the automated workflow.
  • Ingestion points: Design document path provided in the environment or by the user (SKILL.md, Step 1).
  • Boundary markers: Absent; the agent reads the file content directly and passes extracted data to sub-workers.
  • Capability inventory: The skill itself manages workflow state; implementation capabilities are delegated to other specialized skills which interact with the local filesystem and GitHub.
  • Sanitization: None observed; the skill relies on the structural verification of the document (e.g., checking for specific headers like '## User Stories').
Audit Metadata
Risk Level
SAFE
Analyzed
May 15, 2026, 05:43 AM
Security Audit — agent-trust-hub — feature-delivery