issue-start

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

• Third-party content exposure detected (medium risk: 0.30). runtime path issue-start.fetch -> references/issue-collection.md reads outsider-authored GitHub issue body/comments/labels/images (potentially from queue via user-chosen issue_number), then issue-start.classify/plan ingests that free text for analysis.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's runtime installation/fallback steps fetch and install skill code or read raw skill instructions from GitHub (e.g., https://github.com/Mineru98/skills-store/tree/main/.claude/skills/issue-create and the raw.githubusercontent.com SKILL.md links) which are retrieved at runtime and can supply code or procedural instructions the agent will execute or follow.