claude-code-analyzer

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly says it "Searches GitHub for community resources (always enabled)" and includes web_fetch steps to pull community examples/docs which the agent will read and use to create configurations, so untrusted user-generated GitHub/web content can influence actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches live docs at runtime (e.g., https://docs.claude.com/en/docs/claude-code/settings and https://www.anthropic.com/engineering/claude-code-best-practices) and uses that fetched content to generate agent/skill/command configuration and CLAUDE.md, which directly controls agent prompts/instructions and is presented as a required step in the workflow.