mattpocock-dev-workflows

Pass

Audited by Gen Agent Trust Hub on Sep 7, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from codebase files, git diffs, and conversation history to perform its functions.
  • Ingestion points: Conversation history, codebase source files, git diff outputs, and issue tracker content.
  • Boundary markers: None explicitly defined to delimit untrusted data in the prompts.
  • Capability inventory: File creation (ADRs, CONTEXT.md), Git metadata access (diff, log), and publishing to issue trackers.
  • Sanitization: The skill includes a dedicated 'Redact Secrets' phase in the debugging workflow to prevent credentials from being exposed in generated artifacts.
  • [SAFE]: No evidence of malicious command execution, unauthorized network exfiltration, persistence mechanisms, or obfuscation was found. The skill serves purely as a framework for software engineering discipline.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 7, 2026, 07:14 AM
Security Audit — agent-trust-hub — mattpocock-dev-workflows