Give your agent claws
The Clor CLI lets your coding agent create claws, which are background agents that automate anything on a schedule and run on your laptop, Mac mini, or a VM
Claws make AI automation easy
Claws are just agents. They use the same MCP servers, CLIs, skills, and tools your agent does, so anything you can ask your agent to do once, a claw can do in the background or on a schedule.
Examples:
-
Catch new CVEs the morning they hit your dependencies -
Diff competitor pricing pages and flag every material change -
Catch every new SEC filing from companies you track -
Fix broken links across your docs before readers hit them -
Sort inbox receipts into one expense ledger every morning -
Produce a daily audio briefing from your reading list
A toolkit for your agent
Coding agents are already incredibly capable. Clor gives them (and every claw) the tools to work outside the repo, browsing the live web, sending email, moving files between machines, and generating across every major AI provider.
-
Cut token usage by routing each task to the smallest, fastest, or best model that fits -
Ask Gemini, Opus, and GPT to debate, then pick the strongest answer -
Search the web, scrape pages, capture screenshots, crawl whole sites -
Read and send email through any IMAP or SMTP mailbox -
Store files in a cloud drive that follows your agent across machines -
Hand over passwords and API keys through encrypted secrets -
Generate images, speech, and transcripts across GPT, Gemini, and ElevenLabs
Just tell your agent to "create a claw"
Describe what you want and when. Your agent writes the claw and starts it running. Like create a claw to scan Crunchbase News every weekday morning for AI funding rounds over $50M and email me the top 5 by 8am
-
Just describe what you want -
Your agent sets it up and runs it -
Tell your agent to modify, pause, or delete anytime
Start with one of these
Import one and your agent takes it from there
Catch dependency relicensing
-
Watches your dependencies' licenses every weekday for a relicense -
Catches moves from permissive MIT or Apache to restrictive BSL, SSPL, AGPL, or Elastic -
Scrapes the LICENSE file and release notes and explains the commercial impact -
Reports each relicense once, staying silent when nothing changed
Shortlist available domains
-
Generates fresh candidate project and product names from your seed keywords -
Bulk-checks availability with a strict second-pass verify against false positives -
Publishes a living available-names shortlist page behind basic auth -
Emails only when a previously-taken name you wanted becomes newly available
Watch dependency advisories
-
Watches GHSA, OSV, and NVD every morning for advisories hitting your dependencies -
Reads your declared dependencies from a drive manifest, private lockfile optional -
Ranks each advisory by severity with the fixed version and your exposure -
Emails a digest only when a new advisory lands, never the same one twice
Publish a changelog site
-
Turns merged pull requests, release tags, and drive fragments into a polished changelog -
Atomically deploys the site to your clor.app subdomain with rollback safety -
Groups entries by release and rewrites terse commits into readable summaries -
Redeploys only when something new has shipped since the last build
Produce an engineering audio briefing
-
Pulls a drive-managed reading list of engineering blogs and changelogs each weekday -
Summarizes only what is new since the last briefing -
Generates a spoken briefing with ElevenLabs text to speech -
Publishes the audio plus transcript as a private page behind basic auth
Watch vendor status pages
-
Watches your stack vendors' status pages every 15 minutes -
Maps each incident to your surface area and suppresses the irrelevant ones -
Hands over a ready-made customer-facing acknowledgement -
Drops resolved incidents and stays silent when nothing affects you
Cluster support tickets
-
Clusters the day's inbound support mail by underlying issue every morning -
Flags new bugs spreading across customers and release-driven spikes -
Attaches volume and tone to each cluster with a routing read -
Sends nothing on quiet days
Watch public product questions
-
Watches the product subreddit, Stack Overflow tag, and Hacker News every 4 hours -
Surfaces unanswered questions and questions carrying wrong answers -
Ranks by audience reach and your published position on the topic -
Never posts anything; the team chooses what to answer
Summarize known product issues
-
Builds a daily known-issues digest from ticket patterns, vendor status, and the open-issue file -
Updates each ETA from engineering status notes in the drive -
Removes resolved issues automatically so the digest stays current -
Flags what is ready to publish to a public known-issues page