Skip to main content
skill CX Codex official · Anthropic

build-agent · anthropics — Claude Code skill: what it does, how to install

build-agent turns a chore you do by hand into a named skill. You call it by name or run it on a schedule.

Apache-2.0 386 installs ★ 36,654 updated 2026-10-24

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add anthropics/knowledge-work-plugins --skill build-agent -a claude-code
Open on GitHub

What build-agent does

build-agent starts from a task you just did in the conversation, or from your account of the last time you did it. The corrections you made along the way matter most, since they show where your judgment sits. Next it separates what stays fixed, such as order and sources, from what changes each run and from the points where you decide.

Approval gates go where a step sends, publishes, spends or deletes, and where you hesitated. It won't ask before every step. The finished SKILL.md uses your vocabulary: if you say "carrier file", the skill says carrier file. Then it runs the skill on a case whose answer you already know and sets the result beside yours. If they differ, it rewrites the skill and runs it again.

Last, it registers the skill with smb-router so a plain request reaches it, and offers a schedule. For team sharing it spells out who gets the skill and what data it can reach. A shared skill that emails anyone should stay draft-only, because someone who didn't build it doesn't know its assumptions. build-agent won't write a skill that sends or spends without approval gates, whatever you ask.

What it helps with

  • turn a weekly reconciliation into a skill you call by name
  • put a recurring task on a schedule, such as every Monday
  • pull the rules out of a conversation where you corrected the agent as you went
  • place approvals only on irreversible steps and the ones you hesitated over
  • share a skill with the team after agreeing what data it can reach

Example prompts for the agent

  • I build this report by hand every week. Make it a skill I can ask for by name.
  • Use build-agent to turn what we just walked through into a reusable skill.
  • Automate this so it runs every Monday.

How to install build-agent

1Skills CLI (Vercel)

npx skills add anthropics/knowledge-work-plugins --skill build-agent -a claude-code

2Manually: put the files in place

~/.claude/skills/build-agent/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/build-agent/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about build-agent

How do I install build-agent in Claude Code?

Run npx skills add anthropics/knowledge-work-plugins --skill build-agent -a claude-code. The skill file ends up at ~/.claude/skills/build-agent/SKILL.md.

What does build-agent do?

It takes a repeated task you do by hand and writes a skill for it, with approval gates. It then tests the skill on a case you know, registers it with smb-router and offers a schedule.

Can the skills it builds send things without asking me?

No. build-agent does not write skills that send or spend without approval gates. For shared skills that send anything, it recommends draft-only mode.

Source text

SKILL.md · Apache-2.0
---
name: build-agent
description: >
  Turns a task the owner keeps doing by hand into a named, reusable skill they
  can trigger by name or put on a schedule. Watches them do it once or listens
  to them describe it, works out what varies and what stays fixed, writes the
  skill with approval checkpoints in the right places, registers it with the
  router, and shares it with the team. This is how a business gets a workflow
  nobody shipped for them. Use this whenever the owner describes something
  repetitive they wish were automatic — including phrasings like "I do this
  every week," "can you remember how to do this," "make this a thing I can just
  ask for," "build me an agent," "automate this for me," "turn this into a
  workflow," or "I want this to happen every Monday." Reach for it when they
  describe a routine, even without asking for automation.
allowed-tools: Read, WebFetch
---

# Build Agent

Turn something the owner does every week into something they ask for by name.

This is what keeps the plugin from having to ship every industry's workflow. No catalog can cover a septic contractor's permit routine or an insurance agency's carrier commission reconciliation. Those owners can build them.

## Step 1 — Watch it happen, or hear it described

Two starting points, and the first is much better.

**Best: they just did it in this conversation.** If the owner has walked through the task with you — pulled the files, made the decisions, corrected you twice — that transcript is the specification. Extract it rather than re-interviewing them. The corrections they made are the most valuable part, because they encode judgment the owner would never have thought to state.

**Otherwise: they describe it.** Ask them to walk through the last time they did it, concretely. Real specifics beat an abstract description every time — "last Tuesday I pulled the carrier file, matched it against what we'd booked, and chased six discrepancies" tells you more than "I reconcile commissions."

Read `reference/capture.md` for how to get a complete picture without an interrogation.

## Step 2 — Separate what varies from what stays fixed

This is the whole craft of the step. A skill that hardcodes last Tuesday's specifics only works on last Tuesday.

For each part of the task, decide:

- **Fixed** — the sequence, the sources, the output format, the rules
- **Varies** — dates, names, amounts, which file, which customer
- **Judgment** — the parts where the owner decides, which become approval gates

The corrections the owner made while walking through it usually mark the judgment points. When someone says "no, not that one — we skip anything under USD 50," that is a rule. When they say "hmm, depends," that is an approval gate.

## Step 3 — Find the approval gates

Every step that sends, spends, publishes, or deletes needs a gate. So does every step where the owner hesitated.

Do not gate everything — a skill that asks permission nine times is worse than doing it by hand. Gate the consequential and the ambiguous, and let the rest run.

Read `reference/skill_authoring.md` for where gates belong and how to phrase them.

## Step 4 — Write it

Produce a real skill in the same shape as everything else in this plugin: a `SKILL.md` with frontmatter and numbered steps, plus a `reference/` folder if the detail warrants it.

Requirements it must meet, same as every shipped skill:

- Name and folder match, lowercase with hyphens
- Description says what it does **and** when to trigger, in third person
- A real fallback for when a connector is missing
- Approval gates on anything consequential
- Never invents a number; missing data is reported as missing

**Write it in the owner's terms, not in generic business language.** If they call it "the carrier file," the skill says carrier file. A skill full of unfamiliar vocabulary is one they will not trust to run unattended.

## Step 5 — Test it on a real case, before they rely on it

Run the new skill against a case the owner already knows the answer to — ideally the exact one they walked through. Show the output next to what they got by hand.

This is the step that determines whether the skill gets used. An owner who has seen it reproduce a known-good result will schedule it. One who hasn't will run it manually and check it every time, which saves nothing.

If it doesn't match, fix the skill and run it again. Do not ask the owner to accept a near-miss.

## Step 6 — Register and set the cadence

Add it to `smb-router` so plain-English requests reach it, and record the trigger phrases the owner actually uses.

If it is a recurring task, offer to schedule it. Scheduling is what converts "a skill I could run" into "something that happens without me," which is the outcome they wanted.

## Step 7 — Share it, with approval

Offer to share with the team. Say what sharing means: who gets it, what data it can reach, and what it can do unattended.

**Be specific about the risk when a skill sends, spends, or writes.** A skill that emails customers, running for someone who did not build it and does not know its assumptions, is a real hazard. Recommend draft-only for shared skills that send anything.

## What not to do

- **Do not re-interview about a task they just walked through.** The transcript is the spec.
- **Do not hardcode the example.** Names, dates, and amounts vary; the sequence does not.
- **Do not gate every step.** Nine approvals is worse than doing it by hand.
- **Do not skip the test run.** An untested skill gets checked manually forever, which saves nothing.
- **Do not write it in generic business language.** Their words, or they won't trust it.
- **Do not build something that sends or spends without approval gates.** Ever, regardless of what the owner asks for.

## After the build

The owner now has a named skill they can trigger or schedule. If the new skill hit a tool that isn't connected, "connect to my ERP" (`build-connector`) is the natural next step — it turns the missing system into a working source. Also nearby: "brief me" to fold the new output into the daily picture, and "Monday brief" if the skill should feed the weekly one. Offer at most three, and skip any offer the owner already declined this session.

## Reference files

- `reference/capture.md` — getting a complete picture of the task without an interrogation
- `reference/skill_authoring.md` — the structure, frontmatter, and where approval gates belong
- `reference/gotchas.md` — the failure modes that produce a skill nobody uses

© anthropics · Apache-2.0 · anthropics/knowledge-work-plugins

Similar skills

All skills for Automation and no-code: n8n, Zapier, Make →

build-agent by other authors comes first.

build-agent LukeRenton Builds structured Claude Code agent files by asking targeted configuration questions and enforcing formatting rules. CX Codex +1 ★ 333repo stars marketing-loops coreyhaines31 marketing-loops — a skill for Claude Code turns recurring marketing work into scheduled or triggered loops. Each loop defines checks, actions, stored state, and a stop rule. CX Codex +1 59,000installs task-observer Eoghan Henn and contributors task-observer — a skill for Claude Code and five other platforms records user corrections and friction during active sessions. Translates recorded observations into skill updates and new playbooks. CX Codex +1 19,300installs loopify coreyhaines31 loopify turns "this task should run periodically" into a working loop, picking the pacing and adding idempotency plus a bail-out. CX Codex +1 383installs ae-automation ThinkingAIAgenticEngine The ae-automation Claude Code skill creates scheduled automations on the AE platform. It responds to requests for recurring tasks, reports, reminders, or monitoring. CX Codex +1 166installs google-calendar-automation davepoon google-calendar-automation — a skill for creating Google Calendar events, finding open time, and managing attendees through Rube MCP. CX Codex +1 142installs calendar-automation claude-office-skills calendar-automation automates Google Calendar and Outlook workflows: allocates focus time blocks, sends morning Slack briefings, and prepares meeting notes with context. CX Codex +1 6,300installs

From the same set263

The whole set →
ap-processor ap-processor reads vendor bills from the AP inbox or uploaded PDFs and photos, codes them to your accounts and jobs, checks POs, and stages a payment run. Nothing posts without your yes. CX Codex +1 387installs marketing-monday marketing-monday is a Claude Code skill that brings growth metrics, customer feedback, and competitor changes into a weekly brief. It can run once or offer a recurring schedule. CX Codex +1 386installs report-pack report-pack runs scheduled recurring briefings via report-builder and adds context from business-pulse, delivering an XLSX sheet and summary. CX Codex +1 386installs payroll-prep payroll-prep — a skill for Claude Code, Codex, Cursor and other agents that pulls timesheets from Gusto, QuickBooks Payroll, or an uploaded spreadsheet and totals regular, overtime, and PTO hours per employee. CX Codex +1 386installs growth-pulse growth-pulse combines sales, marketing, and funnel metrics from connected services or CSV files into a single report with three weekly growth priorities. CX Codex +1 384installs grow-pipeline grow-pipeline builds a sales pipeline from market research to CRM logging, with ranked prospects, tailored outreach, and approval at each handoff. CX Codex +1 383installs pay-the-bills pay-the-bills processes vendor bills from intake through cash review and a staged payment run. Useful when invoices need coding before anyone approves payment. CX Codex +1 383installs build-connector build-connector finds a working way to link Claude to a tool with no official connector, then carries the connection through to something usable. CX Codex +1 382installs

Project sponsors

Companies that keep this analytics open