pair-agent

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the model to emit the entire pairing instruction block (which contains one-time setup keys/session tokens) verbatim and to accept and run ngrok auth tokens provided by the user, requiring the LLM to handle and output secret values directly.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill's setup step conditionally runs a runtime downloader that does: curl -fsSL "https://bun.sh/install" -o "$tmpfile" && BUN_VERSION=... bash "$tmpfile", which downloads and executes remote code from https://bun.sh/install as part of the required build/setup flow.