antigravity-skill-orchestrator

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

• Third-party content exposure detected (high risk: 0.75). SKILL.md instructs the orchestrator to fetch a master skill catalog from a public URL at runtime (web content), and to read it to select skills—this is outsider-authored free text entering the agent’s LLM context via the runtime web retrieval path.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill explicitly instructs the agent to fetch the master catalog at runtime (https://raw.githubusercontent.com/sickn33/agentic-awesome-skills/main/CATALOG.md) to discover which skills to load, meaning external content fetched during execution can directly alter the agent's behavior.