sync-superpowers

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

• Third-party content exposure detected (high risk: 0.85). SKILL.md describes fetching outsider-authored free text from the public GitHub repo at runtime (e.g., curl/GitHub API calls to obra/superpowers for skills/<skill-name>/SKILL.md and other files), which would be ingested into the agent’s LLM context for classification/absorption.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill issues runtime fetches from the GitHub API and raw content URLs (e.g. "https://raw.githubusercontent.com/obra/superpowers/main/skills//SKILL.md" and "https://api.github.com/repos/obra/superpowers/...") which are parsed and injected into the agent's decision/merge flow, so remote content can directly control prompts and behavior.