Drop-in agent rule

Cursor

Cursor agents can keep CrawlDex as a lightweight route-risk check before they automate or recommend a web action.

Install or paste

.cursor/rules/crawldex.mdc

markdown

---
description: Use CrawlDex before public website tasks
alwaysApply: true
---

When a task involves a public website action or inspection, call CrawlDex before acting.

- Resolve vague user requests with `GET https://crawldex.com/api/v1/resolve?q=...`.
- Preflight the exact site and task with `POST https://crawldex.com/api/v1/preflight`.
- Set `CRAWLDEX_CHANNEL=pack-cursor` or send `x-crawldex-channel: pack-cursor` on CrawlDex calls.
- Treat `decision.recommendation` and `agent_instruction` as binding guardrails.
- Do not submit private, logged-in, payment, account, message, upload, or irreversible data.
- Report only sanitized public outcomes after eligible public work.

Privacy boundary

Keep CrawlDex public-only.

Do not submit private account data, payment data, messages, screenshots with sensitive content, cookies, tokens, raw traces, local files, or logged-in-only pages. Stop before irreversible work unless the user is present and explicitly authorizes it.

Preflight call

bash

curl -sS -X POST https://crawldex.com/api/v1/preflight \
  -H "content-type: application/json" \
  -H "x-crawldex-channel: pack-cursor" \
  -d '{"site":"netflix.com","task":"subscriptions.cancel","agent_profile":{"stack":"browser-agent","browser_runtime":"chromium"}}'

Report redacted outcome

bash

curl -sS -X POST https://crawldex.com/api/v1/runs \
  -H "content-type: application/json" \
  -H "x-crawldex-agent-key: $CRAWLDEX_AGENT_KEY" \
  -H "x-crawldex-channel: pack-cursor" \
  -d '{"site":"netflix.com","task":"subscriptions.cancel","outcome":"success_with_handoff","friction":["login_required"],"source_tier":"anonymous_report","evidence":{"artifact_types":["redacted_agent_report"]}}'