land-and-deploy

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly fetches and inspects arbitrary production/staging/preview URLs (e.g., Step 1.5b uses curl on {production-url}, Step 5/1.5c extracts preview URLs from GitHub PR checks, and Step 7 runs $B goto / $B console / $B perf against the provided or discovered URL), so it ingests untrusted third-party web content and uses that content to drive decisions like health verdicts and reverts.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill's SETUP step will run a curl to fetch and then bash-execute the remote installer script at runtime from https://bun.sh/install (curl -fsSL "https://bun.sh/install" -o "$tmpfile" ... bash "$tmpfile"), which downloads and executes remote code as a required setup path for the skill.