github-sync-helper

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). The skill is a legitimate GitHub helper with no obfuscation, pretext, or credential exfiltration, but it sets undisclosed default commit attribution to a specific third-party account and has a workflow that runs a destructive reset --hard and direct main push without the confirmation it otherwise requires, which is risky to the user without malicious intent.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The gh-issues-list command reads outsider-authored issue titles from the GitHub Issues API at runtime, and any outsider can file an issue on a public repo without the agent selecting a specific item, so this is medium under the GitHub issue rule. gh-pr-list similarly surfaces outsider-authored PR titles.