Agent skill

Ext Develop

by clacky-ai in clacky-ai/openclacky

Build, debug, or publish an OpenClacky extension — scaffold a new one from an idea, fix a broken/invisible panel/api/skill/agent, or ship it to the marketplace.

MITAuto-check passedDevelopment

Install Ext Develop

skills CLI
$ npx skills add clacky-ai/openclacky --skill ext-develop -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install clacky-ai/openclacky ext-develop --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/clacky-ai/openclacky.git skills-src && mkdir -p .claude/skills && cp -r skills-src/lib/clacky/default_extensions/ext-studio/skills/ext-develop .claude/skills/ext-develop && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
ext-develop
GitHub stars
1.2k
Token cost
~3.4k tokens
SKILL.md length
1,787 words
Files
1
Skills in repo
20
Repo updated
First seen
Licence
MIT

At a glance

Build, debug, or publish an OpenClacky extension — scaffold a new one from an idea, fix a broken/invisible panel/api/skill/agent, or ship it to the marketplace.

  • Works in 3 steps: Identify what the user wants and the… → For behavior-only discussion, apply… → Confirm the proposed behavior, affected…
  • Create/start extension
  • SKILL.md covers Before designing or editing, Capability → reference, Working boundaries and Engineering checks, plus 4 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Ext Develop is an agent skill from clacky-ai/openclacky. Build, debug, or publish an OpenClacky extension — scaffold a new one from an idea, fix a broken/invisible panel/api/skill/agent, or ship it to the marketplace. Trigger on create/start extension, plugin, panel, ext verify error, "won't load", "not showing up", publish/ship/unpublish an extension.

Its SKILL.md is about 3.4k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Development. The repository describes itself as: The most Token-efficient open-source AI Agent. The licence is MIT.

When your agent uses it

  • Create/start extension
  • Ext verify error
  • Publish/ship/unpublish an extension

Example prompts

  • “t load”
  • “not showing up”
  • “/ext-develop”

Workflow steps

3 steps, taken from the first numbered list in SKILL.md.

  1. Identify what the user wants and the smallest capability set that can provide it.
  2. For behavior-only discussion, apply Engineering checks without enumerating APIs.
  3. Confirm the proposed behavior, affected files, and visible result with the user

What it can do on your machine

Read from SKILL.md and the folder at commit 66b741e. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    Links to these hosts (documentation or services it may open):

    • openclacky.com

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Ext Develop loads about 3.4k tokens when it runs. Until then it costs about 77 tokens; SKILL.md has 1,787 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~77
When it runs · the whole SKILL.md, loaded when a task matches
~3.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from clacky-ai/openclacky at commit 66b741e, republished under its MIT licence (© clacky-ai). 1,787 words, ~3,423 tokens.

Download SKILL.mdSave it as .claude/skills/ext-develop/SKILL.md (or your agent's skills folder).
name
ext-develop
description
Build, debug, or publish an OpenClacky extension — scaffold a new one from an idea, fix a broken/invisible panel/api/skill/agent, or ship it to the marketplace. Trigger on create/start extension, plugin, panel, ext verify error, "won't load", "not showing up", publish/ship/unpublish an extension.
name_zh
扩展开发
description_zh
开发、调试和发布扩展:从想法搭出新扩展,修复加载失败或不显示的问题,或发布到扩展市场。
agent
ext-developer

Extension Development

Use this skill to guide the workflow, not as a second API manual. The online references below own the detailed contracts, examples, and reload behavior. The local engineering checks below remain mandatory even when those references cannot be fetched.

Before designing or editing

  1. Identify what the user wants and the smallest capability set that can provide it. Check existing host services and shared panels before building a replacement.
  2. For behavior-only discussion, apply Engineering checks without enumerating APIs. Before specifying or implementing an interface, choose its one matching capability reference below; read the overview when container/scaffold/reload behavior is relevant. Reuse verified text still in context. For a named interface, obey the system prompt's strict lookup budget and stop as unverified when it is exhausted. Use the article's contract/examples, not the documentation site's navigation, styles, or scripts. If the user declined a fallback, request only the missing source and never ask again.
  3. Confirm the proposed behavior, affected files, and visible result with the user before scaffolding or editing. Do not publish, restart, or add privileged behavior merely because implementation was approved.
Documentation fallback
  1. Fetch the relevant official URL directly; a web search is not required. For truncated output, read the returned temp_file for the needed section. For HTML pages, web_fetch saves extracted text with whitespace collapsed, not original HTML: line-based reading may still return one long line. One targeted text extraction is enough to check a detail. Do not write helper scripts or try to reconstruct code formatting from that text; lost newlines can change the meaning of code comments and examples. A no-file-change task includes temporary files such as /tmp/*.py: only read the tool-returned cache, do not create a script or reformatted copy elsewhere.
  2. If the needed contract is still missing, truncated, or ambiguous, stop this lookup. Report it as unverified; no match on a page does not prove the method or its namespace absent, private, or unsupported. Do not inventory other APIs to infer an answer. Do not retry extraction or probe other extensions/the running host to fill the gap. Ask permission to open the official documentation in a new browser tab. If already authorized for this task's documentation lookup, do not ask again for each page.
  3. After consent, use the available browser tool to read only the relevant official documentation. Do not navigate, inspect, or close unrelated tabs. Read beyond truncated snapshots as needed; opening a page alone does not verify its contents. If browser setup is required, ask whether the user wants to use browser-setup; never enable it silently or bypass a disabled browser. If the tool is unavailable, do not invent another browser control mechanism or install one without approval.
  4. If the user declines or the browser cannot retrieve the reference, ask the user to supply the documentation or designate a separate source checkout matching the running version and permit read-only inspection. Do not search installed OpenClacky gem source. Reading a checkout does not authorize modifying it to make an extension work.
  5. If still blocked, end with only Unverified detail, Reference checked, and Evidence or permission needed next. For example: "I could not verify this method from this page; please provide its source." Do not turn that into "the object does not exist" followed by a disclaimer. Clarifying the user's goal may continue, but implementation against that contract must wait. Resolve version differences before relying on a contract; version metadata is allowed, installed implementation source is not.

Capability → reference

NeedRead
Container model, source layers, scaffold, reload, verification, publishingExtension System Overview
Manifest metadata, panels, agents, tools, skill restrictions, field validationext.yml Manifest
Slots, agent scope, tab lifecycle/badges, subscriptions, replay, Composer, Aside, Modal, WorkspaceWeb UI Extensions
Extension-owned backend, route parameters, persistent data, session/project helpers, public endpointsHTTP API Extensions
Host sessions/history, files, recovery, billing, media, development UI helpersCalling Host APIs
Agent persona and configurationAgent Configuration
Skill authoring and invocationSkills
IM adapters, optional file delivery and buffered outputChannel Adapters
Tool-call/lifecycle hooks and persistent custom eventsHooks
Explicitly requested runtime patchesRuntime Patches

Read the pages needed for this task, not every reference indiscriminately. This index routes lookups; field lists, signatures, and code examples belong in those pages.

Working boundaries

  • Develop in ~/.clacky/ext/local/<id>/. Never edit installed/builtin packages or gem source to customize a user's extension. Do not search or read installed OpenClacky gem implementation source, even to resolve missing documentation; version metadata is allowed. Confirm intentional id overrides.
  • Do not add hooks or patches without an explicit request; they execute arbitrary Ruby.
  • Reuse the documented host APIs and theme conventions. Do not guess methods from similarly named libraries or manipulate host internals.
  • Keep data outside the extension package using the documented persistence facility. Validate input and get consent for destructive or paid operations.
  • Choose the reload procedure from the overview's contribution-specific table. Do not claim that every change hot-reloads, or restart the server without consent.
  • Verification is evidence, not a prediction: distinguish manifest checks, tests, actual UI/API behavior, and anything still awaiting user verification.
  • Local development does not imply marketplace publication.

Engineering checks

Apply these when proposing the design and again against the actual code before handoff. Do not count a promise to follow a rule as evidence that the implementation follows it.

  1. Access and side effects: identify files, sessions, credentials, and external destinations involved. Refuse unauthorized access or disclosure; never bypass host authentication. Keep secrets out of frontend code and logs. Validate inputs and paths, keep persistent data outside the package, and ensure destructive changes or private-data transfers require specific approval rather than general permission to build.
  2. Request budget: check every fetch, loop, timer, retry, and worker. Prefer events and cached reads; do not repeatedly fetch full histories. Default to no added threads; necessary concurrency must be bounded. No tight loops or sub-second polling. Necessary polling must use a coarse interval, prevent overlap, pause when hidden or finished, and have bounded retries and request duration. Check the failure path as well as success.
  3. Lifecycle and cost: verify rerenders and session switches do not multiply listeners, timers, or requests; disposal releases them and stale responses cannot affect a new session. Hidden cached panels must pause unnecessary work without waiting for disposal. A tab/session switch does not prove disposal occurred; verify unsubscribe/visibility contracts instead of using assumed disposal as a substitute for cleanup. Model/media calls need explicit approval, including recurring scope; mounting, refresh, and history replay must not initiate paid or destructive actions or external sends. Read-only billing queries are not model invocations, but still need request limits.
  4. Host reuse and evidence: reuse host capabilities and btn-* / form-*, Clacky.Modal, and var(--color-*); use scoped, prefixed custom styles where needed. For basic inputs use form-input / form-textarea. Confirm custom property names in the relevant reference or generated scaffold, not by their plausible spelling; do not invent fallback token names. If a token is unverified, omit the color override and inherit host styling, rather than adding a guessed token with hardcoded fallback colors. Prefix tab ids with the extension id to avoid collisions. Check light/dark appearance for UI work. Do not add unrequested hooks or patches or modify installed packages. Test the applicable boundaries above with controlled data; do not use real destructive, disclosure, or paid operations just to prove a safeguard. Report which checks were exercised, which were code review only, and which remain open.
Show full SKILL.md (556 more words)Show less

Scaffold

  1. Agree on the smallest implementation. A panel may reuse a host API without a new backend; a capability composed from existing tools may only need a skill. A user's ordinary app work does not need an extension container, and a project-only skill can live in that project's .clacky/skills/.
  2. Run clacky ext new <id> for a runnable skeleton. Use --full only when the task needs the broader examples. Read generated files before changing them.
  3. Implement the approved behavior against the referenced contracts. Keep manifest paths, unit ids, and actual files aligned; do not retain unused scaffold features.
  4. Adapt the scaffold's tests to the real behavior and run them. If a backend has no tests, add them instead of treating missing tests as permission to skip verification. Cover invalid input and failures as well as the successful response; apply Engineering checks to the implementation.
  5. Run clacky ext verify; resolve errors and review warnings using its actual code, file, and hint. Verify the visible feature or endpoint afterward.
  6. Follow Handoff below.

Debug & verify

  1. Reproduce the reported symptom; inspect the actual local container and relevant reference before proposing a fix.
  2. Run clacky ext verify and read its structured findings. Use clacky ext list to identify the resolved layer and accidental shadowing. Do not maintain a second hardcoded list of accepted keys in this skill.
  3. For invisible panels, check documented slots, tab options, manifest associations, and the browser error stack. For API failures, check the request method/path, handler contract, status, and logs. For stale behavior, check the reload matrix.
  4. Fix the root cause within the agreed scope, rerun relevant tests and verification, then check the reported behavior and the affected Engineering checks. Do not widen scope to unrelated warnings without discussing them with the user.
  5. Follow Handoff below.

Publish (only on explicit request)

  1. Read the overview's current publishing requirements and command options.
  2. Confirm the target local container, intended marketplace status, device binding, and version. Do not bypass authentication, ownership, or encrypted-content limits.
  3. Run tests and verification. If README.md is missing, ask whether to write usage instructions before proceeding; derive them from the actual implemented behavior.
  4. Publish only the approved target/version/status. Updating an existing extension requires manually choosing a greater version in ext.yml; --force does not increment it. Explain failures instead of guessing alternate publish behavior.
  5. Report the actual version and status returned. Unpublishing also needs explicit approval; it is not routine cleanup.

Handoff

  • State what changed, what passed, and what the user still needs to check.
  • After extension edits, use the development UI helpers documented in Calling Host APIs to show the refresh button once; also open the aside for a session-aside panel. Use the injected service/session context, not a guessed host, port, or session id.
  • These HTTP helpers are not browser control. If the user will test the browser, keep that boundary and report UI checks as pending; do not treat it as permission to inspect their tabs. Honor a separate request not to send UI-helper broadcasts.
  • A successful UI-helper response only acknowledges a broadcast. If no matching UI is connected, ask for manual refresh. Verify required Ruby reloads separately; neither UI helper restarts the server.
  • Stop at the approved outcome. Do not publish or perform other release actions as an automatic wrap-up.

© clacky-ai, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in lib/clacky/default_extensions/ext-studio/skills/ext-develop of clacky-ai/openclacky.

Open the folder on GitHubat commit 66b741e

Compare with similar skills

Ext Develop next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Ext Develop compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Ext Develop this skillclacky-ai/openclacky1.2k—~3.4kAutomated safety check: PassMIT
Vercel Composition Patternssupabase/supabase111k59 repos~726Automated safety check: PassMIT
Finishing a Development Branchobra/superpowers296k5 repos~1.9kAutomated safety check: PassMIT
Typescript Advanced Typesrolling-scopes/rsschool-app10k25 repos~4.2kAutomated safety check: PassMPL-2.0
PR Babysitteropeninterpreter/openinterpreter69k3 repos~4.2kAutomated safety check: PassApache-2.0
Code Review ChecklistshareAI-lab/learn-claude-code78k5 repos~1.1kAutomated safety check: PassMIT

Similar skills

  • Official

    React composition patterns that scale. An agent skill from supabase/supabase.

    111k GitHub starsUsed in 59 repos~726 tokens
    DevelopmentAuto-check passed
  • Walks the last step of a branch: confirm tests pass, detect the git environment, ask how to integrate, carry out your choice and clean up the worktree.

    296k GitHub starsUsed in 5 repos~1.9k tokens
    DevelopmentAuto-check passed
  • Typescript Advanced Types

    rolling-scopes/rsschool-app

    Master TypeScript's advanced type system including generics, conditional types, mapped types, template literals, and utility types for building type-safe applications.

    10k GitHub starsUsed in 25 repos~4.2k tokens
    DevelopmentAuto-check passed
  • PR Babysitter

    openinterpreter/openinterpreter

    Watches an open GitHub pull request until it merges, handling review comments, diagnosing CI failures and retrying flaky checks along the way.

    69k GitHub starsUsed in 3 repos~4.2k tokens
    DevelopmentAuto-check passed
  • Code Review Checklist

    shareAI-lab/learn-claude-code

    Reviews code against a five-part checklist covering security, correctness, performance, maintainability and testing, and reports findings in a fixed format.

    78k GitHub starsUsed in 5 repos~1.1k tokens
    DevelopmentAuto-check passed
  • Greploop

    onyx-dot-app/onyx

    Iteratively improves a PR (GitHub), MR (GitLab), or shelved changelist (Perforce) until Greptile gives it a 5/5 confidence score with zero unresolved comments.

    32k GitHub starsUsed in 4 repos~3.3k tokens
    DevelopmentAuto-check passed

More from clacky-ai/openclacky

All 20 skills in this repo
  • Skill Add

    clacky-ai/openclacky

    Install skills from a zip URL or local zip file path. An agent skill from clacky-ai/openclacky.

    1.2k GitHub stars~560 tokensUpdated today
    Auto-check passed
  • Media Gen

    clacky-ai/openclacky

    Generate or edit images, videos, or audio in the current task.

    1.2k GitHub stars~7.3k tokensUpdated today
    Auto-check passed
  • Browser Setup

    clacky-ai/openclacky

    Configure the browser tool for Clacky. An agent skill from clacky-ai/openclacky.

    1.2k GitHub stars~3.1k tokensUpdated today
    Auto-check: notes
  • Cron Task Creator

    clacky-ai/openclacky

    Create, manage, and run scheduled automated tasks (cron jobs) in Clacky.

    1.2k GitHub stars~2k tokensUpdated today
    Auto-check passed
  • Deploy

    clacky-ai/openclacky

    Deploy Rails applications to Railway. An agent skill from clacky-ai/openclacky.

    1.2k GitHub stars~1.9k tokensUpdated today
    Auto-check passed
  • MCP Manager

    clacky-ai/openclacky

    Manage MCP (Model Context Protocol) servers for openclacky: add, list, probe, remove, reconfigure.

    1.2k GitHub stars~3k tokensUpdated today
    Auto-check passed

Categories

Questions about Ext Develop

What does Ext Develop do?

Build, debug, or publish an OpenClacky extension — scaffold a new one from an idea, fix a broken/invisible panel/api/skill/agent, or ship it to the marketplace. Ext Develop is an agent skill from clacky-ai/openclacky. Build, debug, or publish an OpenClacky extension — scaffold a new one from an idea, fix a broken/invisible panel/api/skill/agent, or ship it to the marketplace.

When should I use Ext Develop?

Ext Develop fits situations like: create/start extension; ext verify error; publish/ship/unpublish an extension.

How do I install Ext Develop in Claude Code?

Run `npx skills add clacky-ai/openclacky --skill ext-develop -a claude-code`. Or copy the skill folder (lib/clacky/default_extensions/ext-studio/skills/ext-develop in clacky-ai/openclacky) into .claude/skills/ext-develop in your project. Claude Code loads it when a task matches its description.

How do I install Ext Develop in Codex?

Run `npx skills add clacky-ai/openclacky --skill ext-develop -a codex`. Or copy the skill folder (lib/clacky/default_extensions/ext-studio/skills/ext-develop in clacky-ai/openclacky) into .agents/skills/ext-develop in your project. Codex loads it when a task matches its description.

Can I use Ext Develop in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add clacky-ai/openclacky --skill ext-develop -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/ext-develop, .gemini/skills/ext-develop, .github/skills/ext-develop and .opencode/skills/ext-develop in your project.

What does Ext Develop need to run?

SKILL.md names no scripts, command-line tools or credentials: Ext Develop is instructions for the agent only.

Does Ext Develop access the network?

SKILL.md names 1 domain. As links in the text: openclacky.com. This is read from the text; nothing was executed.

Is Ext Develop safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Ext Develop use?

Ext Develop is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Ext Develop use?

About 3.4k tokens (SKILL.md is roughly 14k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Ext Develop?

Skills that share tags, products or a category with Ext Develop: Vercel Composition Patterns (supabase/supabase, 111k stars), Finishing a Development Branch (obra/superpowers, 296k stars), Typescript Advanced Types (rolling-scopes/rsschool-app, 10k stars) and PR Babysitter (openinterpreter/openinterpreter, 69k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Ext Develop?

clacky-ai (a GitHub organization) maintains it in clacky-ai/openclacky, which has 1,204 GitHub stars. The repository holds 20 skills in this directory. The repository was last updated on October 8, 2026.

Source: clacky-ai/openclacky on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.