narrow-react-prop-types

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

• Third-party content exposure detected (high risk: 0.85). The workflow’s runtime LLM prompt is constructed from GitHub issue/PR comment text (github.event.comment.body) when handling /iterate via bun ci-scripts/agent-iteration.ts --comment-body "$COMMENT_BODY" and then passed into codelayer as --prompt, so outsider-authored free-form comments can be ingested into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 0.90). The GitHub Actions workflow runs "bunx @humanlayer/cli@latest codelayer ..." at runtime, which fetches and executes a remote npm package that directly controls and sends prompts to the Anthropic model (claude-opus-4-8), so it is a runtime external dependency that executes remote code and controls prompts.