skills/mblode/agent-skills/pr-creator/Gen Agent Trust Hub

pr-creator

Pass

Audited by Gen Agent Trust Hub on Nov 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes git and GitHub CLI (gh) commands to manage PRs and commits. It correctly identifies the risk of shell expansion and instructs the agent to use quoted heredocs (<<'EOF') when passing dynamic text to commands to prevent backtick or variable expansion (command injection).
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from repository files (PR templates), commit messages, and branch names. While it lacks explicit boundary markers or 'ignore' instructions for this content, this is a standard operational surface for a PR automation tool and is categorized as low risk.
  • [DATA_EXFILTRATION]: The skill has the capability to upload files to GitHub via the gh pr edit --attach command. This is used for attaching visual assets (images/videos) to pull requests, but could potentially be targeted for exfiltration of sensitive local files if an attacker-controlled prompt manipulates the file path.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 10, 2026, 04:10 AM