serper

Key points before installing/using this skill: - The script requires a Serper API key (SERPER_API_KEY or SERP_API_KEY). The registry metadata incorrectly lists no required env vars — don't trust the registry entry alone; set the key intentionally and securely. - The code auto-loads a .env file located at the skill directory (../.env relative to scripts/search.py). The README also suggests ~/.openclaw/.env, but the provided loader does not read that path — verify where you place your key so it is not accidentally added to a global file or exposed. Prefer a user-scoped or secrets-managed location. - The skill fetches full page bodies from arbitrary result URLs and prints them to stdout as JSON. That behavior is expected for full-content search, but it can leak sensitive content (paywalled/exposed data, PII) into agent logs or downstream processes. Only run this skill in environments where that output is acceptable. - Rate limits and API usage: Serper has quotas; ensure you use a key with appropriate limits and avoid repeated searches — the SKILL.md enforces a 1–2 searches discipline for that reason. - If you need stronger assurance: review the full scripts/search.py file yourself and consider running it in a sandboxed environment first. Confirm the origin of this skill (the registry lists an owner and README points to a GitHub repo); prefer skills with verifiable sources. If you want to proceed, correct the registry metadata to declare the required API key so the platform can surface credential requirements to users, and consider restricting where full-page outputs are stored or sent.