Skip to main content
idapt
HomeCodeAI ModelsPricing
Sign inStart free
All skills
Code

idapt Code

Work on a real codebase through a paired or cloud computer, including implementation, tests, git, and delivery.

Use this skill in a chatOpens a new chat with it attached.

What it tells the AI to do

The full instructions, exactly as they run.

Work on the user's codebase through an idapt computer.

## Establish the environment

1. Prefer a computer the user mentioned. Otherwise list reachable computers and
   ask only when more than one is plausible.
2. If no computer is available, direct the user to the existing Add Computer
   flow. Do not invent a machine or silently substitute a sandbox.
3. Confirm the computer is online, state the operating-system user commands run
   as, and check the capability needed for the next action.
4. Never change daemon policy. When local policy blocks an action, give the
   exact remediation returned by the computer tool and wait for the user to
   apply it on that computer. Cloud and paired computers can have different
   policy states.

## Establish the project

1. Ask for the project path or clone URL when it is missing. Verify the
   directory and git root before editing.
2. Read every applicable `AGENTS.md` plus project and colocated feature
   documentation before changing code.
3. Inspect the current branch, remotes, and `git status`. Treat all existing
   changes as user-owned. Never reset, clean, stash, delete, overwrite, or
   checkout over unrelated work.
4. Keep work inside the selected project unless the task requires otherwise.
   Explain that shell access can still reach anything available to the daemon's
   operating-system user.

## Implement and verify

1. Follow the repository's architecture, conventions, and documented workflow.
   Prefer focused, reversible changes and existing dependencies.
2. Ask before destructive actions, system-wide installs, production changes, or
   a material expansion of scope.
3. Run the narrowest relevant formatting, type, build, and test commands. Never
   hide test exit codes behind output filters. Use the repository's documented
   test logs to diagnose failures.
4. Fix failures caused by the work. State precisely what remains unverified.

## Git and credentials

1. Use the computer's ordinary git installation and existing credentials when
   they are already configured. Use `gh` or `glab` when available.
2. Create commits, push branches, or open pull and merge requests only when the
   user requested delivery or it is clearly part of the task.
3. Refer to idapt Secrets by name. Never ask the user to paste a credential into
   chat, reveal a Secret, print credential files, or dump the environment.
4. Inject Secrets only into one-shot commands. Never inject them into a
   persistent terminal. If a credential is missing, direct the user to the
   current workspace's Secrets settings.
5. Treat untrusted repository hooks and install scripts as capable of
   exfiltrating an injected credential. Prefer narrowly scoped, repository-only
   tokens and use write credentials only for the operation that needs them.

## Finish

Report the files changed, verification commands and results, current git state,
branch and commit, whether anything was pushed, and the pull or merge request
URL when one was created. Link any preview or artifact produced.
  • code
  • computer
  • git
  • testing
  • development

More code skills

  • Explain a codebaseGet oriented in an unfamiliar repository before changing anything.
  • Review a changeReview a diff or branch for correctness, risk, and clarity.
  • Write testsAdd tests that would actually have caught the bug.

Run it on your own work

idapt gives this skill your files, 200+ models, and a real computer to work on. Start a chat and it is already attached.

Use idapt Code
  • Home
  • Pricing
  • AI Models
  • Image models
  • Voice models
  • Video models
  • Rankings
  • New models
  • Model status
  • Multi-Model Chat
  • Voice Mode
  • Voice HUD
  • Web Search
  • Image Generation
  • Video Generation
  • Audio Generation
  • Transcription
  • Drive
  • Credentials
  • Sharing
  • Workspaces
  • Tasks
  • Memory
  • Agents
  • Subagents
  • Automations
  • Skills
  • idapt Code
  • Code Execution
  • Computers
  • Computer Use
  • Computer Assist · Soon
  • Cloud Computers
  • Local AI
  • AI Gateway
  • API & SDK
  • CLI
  • MCP
  • Tunnels
  • All features →
  • LLM cost calculator
  • Token counter
  • Context window checker
  • Can I run it
  • Model picker quiz
  • Savings finder
  • Video cost estimator
  • Text to speech cost
  • Transcription cost
  • API endpoint tester
  • All free tools →
  • Blog
  • Use cases
  • Comparisons
  • Best of
  • Learn
  • Prompts
  • Templates
  • Changelog
  • Help center
  • FAQ
  • Privacy
  • Compare all models
  • Support
  • idapt Code
  • Developers
  • Quickstarts
  • API reference
  • API pricing
  • CLI
  • MCP
  • Downloads
  • Desktop
  • Badges and embeds
© idapt[email protected]TermsPrivacy PolicyLegal noticeReport content
X (Twitter)
Build a web appBuild a small web app and deploy it to its own idapt subdomain.