forge

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 1.00). The GitHub API and release-download URLs are high-risk here because the script auto-fetches prebuilt release assets (executables) from an arbitrary/derived ${repo}, which can be abused to distribute malware; the rustup.rs URL is an official vendor installer and is not flagged.

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

• Third-party content exposure detected (medium risk: 0.30). In the IDE chaining flow described in SKILL.md, the agent reads outsider-authored Jira/GitHub/GitLab ticket JSON (including description/comments and any embedded text) returned by scrutiny forge-fetch and then passes derived brief/context into later LLM agent prompts, so text can come directly from user-submitted ticket content.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's runtime installer scripts call out to the GitHub Releases API (https://api.github.com/repos/${repo}/releases/latest) and download a release binary from GitHub (https://github.com/${repo}/releases/download/v${ver#v}/${asset}) which installs/produces an executable the skill then runs, i.e., fetching and executing remote code at runtime.