The Agent Foundry
Skills · Subagent · Command · PDF
.claude/the-agent-foundry/zip
skills/06
├ command-builder
├ hooks-starter
├ plugin-packager
├ skill-author
├ subagent-scaffold
├ workflow-designer
agents/01
├ skill-reviewer
commands/02
├ /new-agent
├ /new-skill
FIELD-GUIDE.pdf8pp
English
One payment · No account · No subscription
- ✓ Download link in your inbox in seconds
- ✓ Secure card checkout · VAT handled
- ✓ 14-day money back, no questions
- ✓ Every future revision included
- ✓ Runs in Claude Code and claude.ai
Built for Claude first, our recommended agent. Skills follow the open Agent Skills standard (SKILL.md), so they also run in any agent that reads it; subagents and slash commands are Claude Code formats.
The Agent Foundry
Build your own skills, subagents and plugins
six skills·one subagent·two commands·8-page guide
Describe a job you keep briefing by hand and your agent writes the skill for it, reviews the draft, and packages the set as a plugin your team installs in one command.
Executive summary
- The problem
- You brief your AI on the same job every week, and each briefing comes out a little different, so the output drifts with whoever typed it. The method lives in your head, which means nobody can review it and it leaves when you do.
- What it does
- skill-author interviews you and writes the SKILL.md, the skill-reviewer subagent flags vague triggers and invented-number risk, and plugin-packager ships the set so a colleague installs it from a folder.
- Why take it
- The job you repeat is the one worth writing down once. A reviewed skill is method your team can inspect, share and improve, not a prompt someone remembers.
For practitioners with a repeating job nobody has packaged yet, and the teams they share tools with.
In the first week you will
- 01Run /new-skill on the job you keep re-explaining and install the finished SKILL.md before the next session
- 02Hand a draft to the skill-reviewer subagent and get back the vague triggers, the missing output format and the invented-number risk
- 03Package your skills, subagents and commands with plugin-packager so a colleague installs the whole set from a folder
Contents of the Kit
- 01Six installable agent skills: skill-author, subagent-scaffold, command-builder, hooks-starter, plugin-packager, workflow-designer
- 02The skill-reviewer subagent: reads a draft skill and reports vague triggers, missing output formats, invented-number risk and prompt-injection surface
- 03/new-skill and /new-agent slash commands
- 04Templates for a skill, a subagent, a command, a plugin manifest and a hooks file, in the current Claude Code formats
- 05The Connector Registry: twelve connectors and MCP servers for analytics, search, ads, CRM, database, spreadsheets, browser and files, each with the read-only-first rule
- 06An 8-page Field Guide on the delegation ladder and where each tool belongs on it, and plugin manifest included
One skill, as it ships
SKILL.md
skill-author
- Use when
- Use when the user says "make this a skill", "I keep explaining this", "turn my prompt into a skill", or asks how to write a SKILL.md.
- Requires
- A description of the repeating job, in the user's words
- Will not
- Never write a skill for a job the user has not done by hand at least once · No em dashes in the file · Do not add frontmatter fields the docs do not define
- Returns
- {Title}
One of six skills in the kit · All of them install the same way
The contents, in full · six skills, one subagent, two commands
- 01skill-authorAsks about the inputs, the steps, the output and the checks a human does, then writes a complete skill in the standard format and reviews it for ambiguity and missing refusals.
- 02subagent-scaffoldDesigns a subagent: the job, the minimal tool set, the model, the boundary, the shape of its report back. Writes the file.
- 03command-builderTurns a frequent request into a slash command with argument hints, pointing at the skill that does the work.
- 04hooks-starterA settings hooks block for the guardrails most teams want first: block edits to secret files, run a formatter after edits, warn before destructive git commands, log tool use.
- 05plugin-packagerSkills, agents, commands and hooks into a plugin: the manifest, the folder layout, a local marketplace file for installing from a folder, versioning, a release checklist.
- 06workflow-designerDecomposes a job into a chain, places each step on the delegation ladder (autocomplete, draft, execute with review, execute autonomously, agent), and specifies the verification at each handoff.
- 07skill-reviewer (subagent)The second reader every skill should have before it goes into a team folder. It reads a draft and reports vague triggers, missing output formats, invented-number risk and prompt-injection surface.
- 08/new-skillWrites a new SKILL.md for the repeating job described in the arguments, then offers the skill-reviewer pass on the draft.
- 09/new-agentDesigns a subagent for the job described in the arguments and writes its file, with the install path, the invocation line and one test task.
- 10Connector RegistryTwelve connections that give the model read access to where your work already lives: analytics, search, ads, CRM, database, spreadsheets, browser and files, each with install pointers and the rule that keeps connection safe.
Every other kit in this library was built the way this one teaches: interview the job, write the skill, review it, package the set. The kit is written for people who are not programmers. The formats are the current Claude Code formats, checked against the documentation, and each template is a file you can fill rather than a concept you have to understand first.
Pairs with
The Augmented Practitioner explains what a skill is and why it beats re-prompting. The Agent Foundry is where you build the next one.
Before you buy
01Do I need Claude Code?
No. Every skill uploads to claude.ai as a zip. Subagents and slash commands are the Claude Code extras.
02I am not a programmer.
Install is copying folders into a .claude directory. The Field Guide in each kit assumes nothing beyond that.
03Will it work in my agent?
Skills follow the open Agent Skills standard (SKILL.md), so any agent that reads it can run them. Built for Claude first, our recommended agent.
04Do the tools invent numbers?
Never. Every figure is supplied by you, confirmed by you, or labeled as an assumption in the output.
05How does it arrive?
Instantly. Pay by card and the download link is in your inbox within seconds, and on the confirmation screen.
06Is this a subscription?
No. One payment, lifetime access to your download, every future revision of what you bought included.
07What if it is not useful to me?
Write to info@conclusionmp.com within 14 days of purchase and you get a full refund. No questions, no forms.
Ready when you are
The Agent Foundry, $12.
Instant download, one payment, every future revision included. Or take The AI Shelf, six items for $39, and this kit works out at $6.50.