skills/pubx-ai/skills/local-pr-review/Gen Agent Trust Hub

local-pr-review

Pass

Audited by Gen Agent Trust Hub on Aug 4, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes standard version control commands such as git diff, git log, and git fetch to retrieve branch history and evaluate code changes against a base branch.
  • [COMMAND_EXECUTION]: It utilizes the GitHub CLI (gh) to resolve repository metadata, specifically identifying the default branch and recent pull request targets to determine the correct review base.
  • [EXTERNAL_DOWNLOADS]: The skill synchronizes with the remote repository using git fetch origin to ensure the local tracking references are current before performing comparisons.
  • [PROMPT_INJECTION]: The skill processes untrusted data which creates an indirect prompt injection surface.
  • Ingestion points: Reads git diff output and full file contents from the current branch (SKILL.md, Step 2 & 3).
  • Boundary markers: Instructions require a specific output contract and subagent dispatch prompt to focus analysis, though no explicit 'ignore instructions in code' warnings are provided.
  • Capability inventory: Uses git and gh for data retrieval and provides instructions for subagent spawning.
  • Sanitization: No content sanitization or escaping is mentioned before processing code diffs.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 4, 2026, 02:21 PM
Security Audit — agent-trust-hub — local-pr-review