pr-review-html

Pass

Audited by Gen Agent Trust Hub on Aug 10, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes local commands using the GitHub CLI (gh), a Python script located at ~/.claude/scripts/build_review_html.py, and a local binary named pulsar. These operations are consistent with the skill's purpose as a developer workflow automation tool.
  • [DATA_EXFILTRATION]: The skill generates review artifacts and moves them to $HOME/CodeReviews/ via the pulsar publish command. This represents local data organization within the user's home directory rather than external exfiltration.
  • [PROMPT_INJECTION]: The skill processes untrusted data from GitHub PR bodies and code diffs. While this presents an indirect prompt injection surface, the skill implements safety measures by specifying that the reporting script should HTML-escape fields to prevent cross-site scripting (XSS) in the generated review page.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 10, 2026, 02:41 AM
Security Audit — agent-trust-hub — pr-review-html