Menu
Detects whether the current page contains a captcha challenge (reCAPTCHA, hCaptcha, Cloudflare Turnstile, GeeTest, etc.). Reports detection and stops. Never attempts to solve. See docs/ethics.md.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Signs into a site using credentials provided via `vars` (typically sourced from environment variables) or via a pre-saved Playwright persistent context. Never harvests or stores credentials. See docs/ethics.md.
Extract structured data from an HTML table, optionally walking through paginated views. Returns a list of row dicts keyed by header cells. The demo skill (used to monitor conference paper lists).
Confirms that the current page has finished loading and is ready for interaction. Detects loading spinners, skeleton screens, and content placeholders. Run before extraction skills.
Fills a multi-step form (stepper wizard) by routing labeled fields to values from `vars`, clicking "Next" between steps, and submitting at the end. Pairs with the `searchable-dropdown` and `date-picker-widget` skills for compound widgets.
Uploads a local file to an HTML5 file input, or triggers a download and captures the resulting path. Composes with multi-step-form for file-attaching workflows.
Picks a date in a calendar-style date-picker widget — clicks the input, navigates months if needed, clicks the target day. The famous "vision agent failure case" that motivates this whole library.
| name | detect-captcha |
| description | Detects whether the current page contains a captcha challenge (reCAPTCHA, hCaptcha, Cloudflare Turnstile, GeeTest, etc.). Reports detection and stops. Never attempts to solve. See docs/ethics.md. |
| version | 0.1.0 |
| allowed-tools | ["wait_for_dom_ready","extract_text","assert"] |
| metadata | {"applies_to":"any-website","url_patterns":["*"],"dom_markers":[".g-recaptcha",".h-captcha",".cf-turnstile","iframe[src*='recaptcha']","iframe[src*='hcaptcha']","iframe[src*='challenges.cloudflare.com']"],"flake_rate_target":0.02,"flake_rate_measured":null,"exercised_on":["cloudflare-protected-test"],"cost_budget":{"deterministic_only":true,"max_vision_calls":0},"sensitive":false} |
Run any time the deterministic recipe of another skill fails on what looks like it should have worked, OR proactively before any step that involves form submission, login, or rate-sensitive interaction. The matcher runs this when known captcha markers appear in the DOM.
This skill is detect-only. It identifies captcha provider, surfaces metadata, and stops. If you need to bypass a captcha, you are out of scope of this project — see docs/ethics.md. PRs adding solving logic will be closed.
The skill always returns success if the recipe completes. The interesting state is in the returned extracted map:
$captcha_marker populated → captcha detected (the value identifies the provider via its container class or iframe URL)$captcha_marker not populated → no captcha presentThe caller decides what to do next: stop, prompt the user, switch to a non-captcha-protected URL, etc.
.grecaptcha-badge selector covers this case; it's in v0.2.0.cost_budget.max_vision_calls: 0).Agent task: "Search for foo on site X and grab the first 3 results."
→ invoke_skill("verify-page-loaded") → success
→ invoke_skill("dismiss-cookie-banner") → success
→ invoke_skill("detect-captcha") → success, $captcha_marker = None
→ invoke_skill("search-and-filter") → success
...vs...
→ invoke_skill("detect-captcha") → success, $captcha_marker = "<reCAPTCHA iframe>"
→ agent: "Site X is showing a reCAPTCHA. I cannot continue automatically. Switch to a different source?"
This skill exists to make captcha presence visible to the agent so the agent (and the user) can make an informed decision. The decision is not "now solve it." md](../../DECISIONS.md).
verify-page-loaded — captchas often render after page load; run loaded-check firstdetect-rate-limit (v0.2.0) — captchas often appear after rate-limit triggers