design-plan

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

• Third-party content exposure detected (high risk: 0.75). The skill’s runtime workflow may fetch a referenced GitHub issue via gh issue view ... --json ...body..., and the issue body is outsider-authored free text that can be ingested into the agent’s LLM context during Phase 1.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill explicitly instructs runtime fetching of a GitHub issue via the command "gh issue view {number} ..." (i.e., the remote GitHub issue URL like https://github.com/.../issues/{number}), which will pull external issue content into the agent's context and can directly control the agent's prompts/behavior.