peach-check

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill instructs collecting raw job logs and passing them inline to agents, then explicitly asks agents to return key log line(s) as Evidence (up to 2 lines), which can contain API keys or secrets and thus requires the LLM to handle/output secret values verbatim.

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

• Third-party content exposure detected (high risk: 1.00). This skill explicitly fetches GitHub Actions run/job data and downloads job logs via gh api (e.g., repos/SonarSource/peachee-js/actions/runs/... and repos/.../actions/jobs/JOB_ID/logs) and passes those user-generated logs inline to per-job agents to read and classify, so untrusted third-party content can directly influence agent decisions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill invokes GitHub API endpoints at runtime (e.g., repos/SonarSource/peachee-js/actions/jobs/JOB_ID/logs) and injects the fetched job logs directly into per-job agent prompts, so external content fetched during execution can directly control the agents' input.