clawdtributor

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

• Third-party content exposure detected (high risk: 1.00). The skill's "Live GitHub recheck" and "Review depth" steps explicitly use gh api calls to fetch PRs/issues from the public GitHub repo (repos/openclaw/openclaw) and instruct the agent to inspect PR bodies, comments, and files, meaning untrusted, user-generated third-party content is read and used to decide actions (e.g., listing, closing, or commenting).

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill issues runtime GitHub API calls (e.g., gh api repos/openclaw/openclaw/pulls/<number> and gh api repos/openclaw/openclaw/issues/<number>) to fetch PR/issue bodies and comments which are then injected into the agent's review context and can directly influence prompts/decisions, so this is a runtime external content dependency that can control the agent.