curlshot
Skill21 starsChanged 5 months agoMITView on GitHub
Captures web pages as PNG, JPEG or WebP screenshots, PDFs, or scrolling videos through the CurlShot screenshot API, and renders raw HTML or Markdown to an image or PDF. Use when a task needs a screenshot, full-page capture, mobile or dark-mode view, PDF, or video of a public URL, or an image of generated HTML. Needs a CurlShot access key in CURLSHOT_API_KEY.
- Reads credentials
- Sends data out
1 other project uses a version of this, 1 of them edited.
Adopted by
Most recently taken up by
- SergiuPogor/curlshot-mcpedited, first seen Oct 17, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 1
- SergiuPogor/curlshot-mcp★ 015% same
Wrote their own version · 0
None found.
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository35 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- api-and-interface-designSkill
- ax-extract-workflowSkill
- browser-testing-with-devtoolsSkill
- ci-cd-and-automationSkill
- code-review-and-qualitySkill
- code-simplificationSkill
- context-engineeringSkill
- debugging-and-error-recoverySkill
- deprecation-and-migrationSkill
- documentation-and-adrsSkill
- doubt-driven-developmentSkill
- eval-driven-developmentSkill
- frontend-ui-engineeringSkill
- git-workflow-and-versioningSkill
- hermes-tweetSkill
- idea-refineSkill
- incremental-implementationSkill
- interview-meSkill
- linkedinSkill
- loootSkill
- mcp-tool-integrationSkill
- observability-and-instrumentationSkill
- performance-optimizationSkill
- plan-ledger-tasks-yyloSkill
- planning-and-task-breakdownSkill
- security-and-hardeningSkill
- shipping-and-launchSkill
- source-driven-developmentSkill
- spec-driven-developmentSkill
- test-driven-developmentSkill
- using-awesome-ai-skillsSkill
- video-routerSkill
- xquik-public-dataSkill
The file
We split this skill into the rules it gives an agent
10 rules. About 0 are common, 6 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Use when the user wants a screenshot or full-page capture of a public URL.
- Distinctive: line 15
Use when the user wants a page as it looks on a phone or tablet preset, at a given window size, or in dark mode.
- Distinctive: line 16
Use when the user wants a page saved as a PDF or as a scrolling video.
- Distinctive: line 17
Use when generated HTML or Markdown (a social card, a report, a local page) must be rendered to an image or PDF.
- Distinctive: line 18
Do not use for localhost or private network URLs by address: the API refuses them.
- Very specific: line 22
Check the key without printing it: test -n "$CURLSHOT_API_KEY".
- Very specific: line 33
Add options as extra -d name=value pairs: viewport_device=<preset> (list: https://curlshot.com/docs/devices.md), …
- Very specific: line 44
Use "markdown": "..." instead of html for Markdown, and "format": "pdf" for a PDF.
- Very specific: line 45
Check the result. HTTP 200 with an image/*, video/* or application/pdf content type means success.
- Distinctive: line 49
Every code: https://curlshot.com/docs/errors.md
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
CurlShot Screenshots
Overview
CurlShot renders a page in a real browser and returns the result as an image (PNG, JPEG, WebP), a PDF, or a video of the page scrolling (MP4, WebM, GIF). One HTTP request does it: GET https://curlshot.com/api/v1/screenshot?url=… for a public URL, or a POST with a JSON body for raw HTML or Markdown. Agents with MCP support can use the hosted server https://curlshot.com/api/mcp (tools take_screenshot, list_devices) instead of curl.
When to Use
- Use when the user wants a screenshot or full-page capture of a public URL.
- Use when the user wants a page as it looks on a phone or tablet preset, at a given window size, or in dark mode.
- Use when the user wants a page saved as a PDF or as a scrolling video.
- Use when generated HTML or Markdown (a social card, a report, a local page) must be rendered to an image or PDF.
- Do not use for localhost or private network URLs by address: the API refuses them. Send the page's HTML instead.
Process
- Check the key without printing it:
test -n "$CURLSHOT_API_KEY". If it is missing, ask the user for one. A free key (100 screenshots a month, no card) comes from https://curlshot.com/register. Capture a URL. Send the key in the
X-Access-Keyheader, never in a link you show the user:curl -sS -G "https://curlshot.com/api/v1/screenshot" \ -H "X-Access-Key: $CURLSHOT_API_KEY" \ --data-urlencode "url=https://example.com" \ -d format=png -d full_page=true -d block_cookie_banners=true \ -o shot.png -w '%{http_code}\n'- Add options as extra
-d name=valuepairs:viewport_device=<preset>(list: https://curlshot.com/docs/devices.md),viewport_width/viewport_height,dark_mode=true,selector=<css>,format=jpeg|webp|pdf|mp4|webm|gif,block_ads=true,wait_until=networkidle,delay=<seconds>,cache=true. Every option: https://curlshot.com/docs/options.md Render raw HTML or Markdown with a POST:
curl -sS -X POST "https://curlshot.com/api/v1/screenshot" \ -H "X-Access-Key: $CURLSHOT_API_KEY" \ -H "Content-Type: application/json" \ -d '{"html": "<h1>Hello</h1>", "format": "png", "viewport_width": 1200, "viewport_height": 630}' \ -o card.pngUse
"markdown": "..."instead ofhtmlfor Markdown, and"format": "pdf"for a PDF.- Check the result. HTTP
200with animage/*,video/*orapplication/pdfcontent type means success. Any status of400or above returns a JSON error body. Branch on itserror_code:400/413/422(invalid_options,invalid_url,selector_not_found, ...): do not retry. Fix the request using theerrorsfield.401: the key is missing or invalid.402 quota_exceeded: tell the user and showupgrade_url.403 host_not_allowed: send HTML instead of a private URL.429: waitRetry-Afterseconds, then retry once.5xx: retry once. Fortimeout, dropwait_until=networkidleor lowerdelay. Every code: https://curlshot.com/docs/errors.md
Common Rationalizations
| Rationalization | Reality |
| "curl exited 0, so the file is an image." | A failed render writes a JSON error into the output file. Check the HTTP status and content type first. |
| "Putting the key in the URL is simpler." | Links get logged and shared. Use the X-Access-Key header. |
| "Retrying will fix a 400." | 4xx errors are deterministic. Read errors and fix the request. |
"localhost will work if I try again." | Private addresses are always refused. POST the HTML instead. |
| "A misspelled option will just be ignored." | Unknown options are rejected with invalid_options and a hint. Use the suggested name. |
Red Flags
- The access key appears in output, a committed file, or a URL shown to the user.
- The saved
.pngor.pdfis a few hundred bytes of JSON. - A loop retries on
4xxerrors or ignoresRetry-After. - The capture targets a private address (localhost, 192.168.x.x) by URL instead of sending the page's HTML.
Verification
CURLSHOT_API_KEYwas checked without printing its value.- The HTTP status was
200and the content type matched the requested format. - The output file opens as the expected format and was saved where the user asked.
- Any error was handled according to its
error_code, with at most one retry for429or5xx.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

