skills/steipete/openclaw/gh-issues/Gen Agent Trust Hub

gh-issues

Pass

Audited by Gen Agent Trust Hub on Jul 13, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection. It fetches external content (GitHub issue titles, bodies, and review comments) and interpolates them directly into prompts for sub-agents. These sub-agents have significant capabilities, including shell access and the ability to perform network operations, which could be abused if an attacker-controlled issue contains malicious instructions.
  • [COMMAND_EXECUTION]: The orchestrator and sub-agents execute shell commands such as git, curl, and jq. Additionally, the skill uses node -e to dynamically execute inline JavaScript for parsing configuration files and extracting tokens.
  • [CREDENTIALS_UNSAFE]: The skill programmatically reads GitHub authentication tokens from local configuration files located at ~/.openclaw/openclaw.json and /data/.clawdbot/openclaw.json. While necessary for functionality, this pattern involves direct access to sensitive credential stores.
  • [EXTERNAL_DOWNLOADS]: The skill performs extensive network operations targeting the GitHub REST API (api.github.com) to fetch issues, read reviews, and create pull requests.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 13, 2026, 02:11 PM
Security Audit — agent-trust-hub — gh-issues