claw-shell-hardened

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

• Third-party content exposure detected (high risk: 1.00). The skill's runtime (handler.js sendCommand/readOutput) executes arbitrary shell commands and returns their output to the agent, and SKILL.md even lists examples like "bird read https://x.com/..." — meaning the agent can fetch and ingest untrusted public web/social content that could contain instructions and materially influence subsequent actions.