Skip to content

the2hourclo/ai-employee-builder

v0.27.0UNLICENSED

Connect AI Employee Builder to Codex, Claude Code, Cowork, and other Agent Plugins-compatible hosts.

ai-employee-map

Builder employee · The free introductory AI Employee Map helps owners assess AI readiness across four levels (Manual → Documented → AI as a Tool → AI Employees) and identify a first AI opportunity. USE WHEN user explicitly says "ai employee map", "find AI opportunities", "where can I use AI", "ai readiness", "what level is my business", "am I ready for AI", or asks for the free intro map. Operational business diagnosis routes to business-x-ray; post-map build guidance routes to roadmap.

behavioral-science

Writing employee · Apply behavioral science principles from Chase Hughes' Behavior Ops Manual to any content. Diagnose missed persuasion triggers, enhance drafts with specific techniques, or look up principles by category. 290 principles across 10 categories. USE WHEN user says 'behavioral science', 'behavior ops', 'apply behavioral lens', 'check persuasion', 'enhance with psychology', 'what behavioral principle', 'why isn't this landing', or when any content needs a behavioral science audit.

business-os

Builder employee · Install and maintain the file system AI employees navigate. Builds a BUSINESS-MAP.md world model; department folders with neutral DEPARTMENT.md sources and generated CLAUDE.md plus AGENTS.md adapters; client records; shared system records for CRM, email, Drive, Notion, and other tools; a normalized client-system relationship ledger with backlinks instead of duplicated tool setup; live-data routing with verified IDs; safe migration, tidy, and routing verification. USE WHEN user says "set up my business OS", "organize my workspace", "map my clients", "connect my CRM", "connect my data sources", "make this work in Claude and Codex", "AGENTS.md", "CLAUDE.md", "my AI can't find anything", or "where does my live data live". Do NOT use to diagnose business processes or draw opportunity maps; use business-x-ray for that.

business-x-ray

Builder employee · Operational Business X-Ray: map, diagnose, and optimize how the business runs through progressive visual analysis. Creates Business Maps, Bow-Tie Funnels, Process Swimlanes, a System Connection Map, and a future-state Operating System Map of connected AI Employees and tools. ALWAYS USE for 'business x-ray', 'x-ray my business', 'map my business', 'diagnose my business', 'audit my business', 'find the bottlenecks', 'where am I the bottleneck', 'map my operations', 'where does my business break without me', 'score my business assets', 'leverage score', or a visual operational diagnosis. Folder/file scaffolding routes to business-os; an explicitly requested free intro AI-readiness map routes to ai-employee-map; post-diagnosis build guidance routes to roadmap.

check-setup

Setup employee · Verify and repair this AI Employee Builder setup — connector, license plan, workspace folders, onboarding state, and the managed CLAUDE.md block — then offer consent-gated fixes. USE WHEN user says 'check my setup', 'verify my setup', 'is my setup ok', 'is everything connected', 'something seems broken', 'my skills stopped working', 'fix my setup', 'setup health check', 'am I connected', 'why isn't this working', or another skill or message told them to run a setup check. Safe to run any time — it changes nothing without an explicit yes.

contextual-insights-cta

Writing employee · Write CTAs that continue the conversation instead of interrupting it. Three-element structure (concept link, curiosity bridge, soft CTA) that makes the offer feel inevitable, plus the 7-Element Close checklist (Bridge, Value Stack, Price Reveal, Risk Reversal, Objection Handling, Urgency, Decisive Action) for full paid-offer closes. USE WHEN user says 'write CTA', 'improve this CTA', 'CTA feels generic', 'make the CTA better', 'contextual CTA', 'the offer feels plugged in', 'write the close', 'check the close', 'closing section', '7 elements of a close'.

copywriter

Writing employee · Conversion copywriting for offers, sales pages, landing pages, campaigns, email sequences, welcome/onboarding emails, product bullet points, product benefits, sales bullets, VSLs, workshops, and positioning. ALWAYS USE instead of write for 'write bullet points for my product benefits', 'product benefits', 'sales bullets', 'welcome email', 'email sequence', 'sales page', 'landing page copy', 'campaign', 'offer document', 'VSL', 'workshop landing page', 'what should I sell', 'roast my offer', or 'clarify my offer'. Includes campaign blueprint through deployment plus standalone workflows for welcome emails, Product Bullet Points, Agreement Analogy, Offer Clarity, Game Worth Playing, VSL, Workshop Landing Page, Three-Pillar Offer, Rulebook, and offer QA. Do NOT use for a general newsletter, article, daily email, or broadcast email; those route to write.

decision-tree

Builder employee · Navigate a real branching taste or judgment decision instead of averaging options: which avatar, angle, positioning, hook, offer frame, or thumbnail concept. USE WHEN user says 'build a decision tree', 'map the options', 'map the decision', 'which option is best', 'help me decide' for a creative choice, or when a chain reaches a fork with two or more legitimate branches. Record each human pick and reason, then pass only the chosen branch downstream. Do NOT use when there is no genuine fork among alternatives.

define-brand

Writing employee · Define or sharpen a personal/business brand through a guided interview, then produce two working references — a one-page positioning spine and a False Beliefs Document. Runs the brand questions one or two at a time, open-question-first then a proposed draft to react to, and holds every answer against the branding principles. USE WHEN the user says 'define my brand', 'brand worksheet', 'sharpen my positioning', 'who am I for', 'what's my pairing', 'what do I stand against', 'find my unique edge', 'work on my brand', or BEFORE creating content when no brand is defined yet. Do NOT use to write a finished piece (that's the content skills), to set up or calibrate voice files/voice samples (that's setup-content-employees), to map operations (that's business-x-ray), or to design visuals (that's visual-content).

excali-graphic

Writing employee · Render legible hand-drawn 'Excalidraw whiteboard' diagrams, graphs, and charts as PNGs from hand-authored HTML/SVG — white background, real Excalifont, an SVG wobble filter on SHAPES ONLY so text stays crisp, color-by-meaning accents. The go-to recipe for email graphs, newsletter diagrams, process maps, and video teaching slides. USE WHEN user says 'make a graph for my email', 'excalidraw diagram', 'whiteboard graphic', 'hand-drawn diagram', 'diagram this', 'make a chart', 'draw a process diagram', 'visualize this as a graph', 'teaching-slide graphic', 'sketch this concept on a board', or wants a clean hand-drawn diagram/graph for a video slide, email, newsletter, carousel slide, or canvas. Do NOT use for: photoreal/AI-generated imagery (outside this plugin's scope), or a full multi-slide social carousel (→ carousel — though its individual diagram slides can be drawn HERE).

install-skill

Setup employee · Import and install skills from external sources into your workspace. USE WHEN user says 'install this skill', 'import a skill', 'add this skill pack', 'install from github', 'set up this skill', 'load skill from', 'bring in this skill'.

invitation-lens

Writing employee · The selling philosophy beneath every CTA, offer, and promotional line — invite, don't push; recognition, not manipulation — plus the premium framing shifts (reserve not "lock in", included resource not "bonus"). USE WHEN user says 'apply invitation lens', 'make this less salesy', 'this feels pushy', 'soften the sell', 'invitation framing', 'invite don't sell', 'less aggressive CTA', 'sounds too salesy', or when any selling/persuasion copy needs its relationship-to-the-reader checked. Do NOT use to WRITE a CTA from scratch (→ contextual-insights-cta) or a hook (→ write-hook) — this is the LENS those skills apply.

meta-create-agent

Builder employee · Create subagents for Claude Code. USE WHEN user says 'create an agent', 'make an agent', 'add an agent', 'new agent for', 'build an agent', 'meta create agent', 'I need a specialist for', 'delegate this to', 'autonomous worker for'.

meta-create-command

Builder employee · Create slash commands for Claude Code. USE WHEN user says 'create a command', 'make a command', 'add a command', 'new slash command', 'build a command for', 'I need a shortcut for', 'make a quick action', 'automate this action'.

meta-create-hook

Builder employee · Create hooks for Claude Code automation. Covers ALL 14 hook events, ALL 6 hook locations (settings.json, frontmatter, plugins), ALL 3 hook types (prompt, command, agent), async hooks, input modification, and decision control. USE WHEN user says 'create a hook', 'add a hook', 'meta create hook', 'automate when', 'trigger when', 'run automatically', 'after writing', 'before commit', 'validate output', 'add automation to skill'.

meta-create-mcp

Builder employee · Add, configure, or build MCP servers in Claude Code. USE WHEN user says 'add an MCP', 'connect to', 'add integration', 'setup MCP', 'connect MCP server', 'add notion', 'add youtube', 'configure MCP', 'build an MCP', 'create MCP server'.

meta-create-plugin

Builder employee · Package skills into a Claude Code plugin for distribution. Creates plugin folder structure, plugin.json manifest, and prepares for publishing. USE WHEN user says 'create a plugin', 'package these skills', 'bundle skills into a plugin', 'make a plugin from', 'publish plugin', 'plugin from skills'.

meta-create-skill

Builder employee · Create, validate, test, update, and harden agent skills for Claude or Codex; chain them into Skill Systems (AI Employees); choose the artifact an outcome needs; and apply domain-invariant contracts to stateful or side-effecting employees. Use when someone describes an OUTCOME ('review emails weekly', 'I do this manually'); asks to build ('create a skill', 'build an AI employee', 'automate this operation', 'make it decide like me'); asks to change an existing skill ('update skill X', 'add a rule'—always load this first); asks to protect business state ('add ontology', 'enforce invariants', 'stop duplicate or illegal actions'); reports symptoms ('it ignores its rule', 'works sometimes', 'said it did it but did not', 'make it production-grade'); or wants to schedule, validate, test, or audit skills. Outcome-only requests route to artifact selection first. A skill that MISFIRED and needs a patch routes to retrospective.

narrative-gym

Writing employee · Apply the ABT (And, But, Therefore) narrative framework as a writing lens. Builds narrative backbone before writing or diagnoses why existing content isn't landing. USE WHEN user says 'apply abt', 'narrative lens', 'build my abt', 'find my hook', 'find my but', 'what's my tension', 'analyze my content', 'check my narrative', 'is this narrative', 'strengthen my hook', 'narrative gym', 'abt framework', or before starting any content piece that needs a clear hook and structure.

onboard

Setup employee · First-run setup for the ai-employee-builder plugin. Scaffolds the authoring folders (.claude/skills|agents|commands|hooks), writes a workspace CLAUDE.md with skill routing, and installs the Business OS (department folders + BUSINESS-MAP.md) — then hands off to the tracked build journey, whose first checkpoint is mapping the business with the Business X-Ray. USE WHEN user says "onboard me", "set me up", "get me started", "start my workspace", "bootstrap my workspace", "first time setup", "build my workspace", OR when you detect a fresh workspace (no .claude-state/onboarding-progress.json) and the user has just installed the ai-employee-builder plugin.

perspective-shift

Writing employee · The shared brain of the Perspective Shifting Engine — used by ANY short-form format (reel, carousel, thread, quote post, email opener), not reel-bound. Finds the ONE charged perspective shift inside any source (transcript, article, newsletter, raw idea, or your brand-positioning doc) and develops it into a scoped, fully-charged shift. Runs source × your brand-positioning doc to produce charged shift cards (NOT generic summaries), picks the standalone shift + its content-type angle, and builds its dream/nightmare stakes ladder. Format skills (carousel, thread, reel) call this for their idea/shift step. USE WHEN user says 'find the perspective shift', 'what's the shift here', 'extract the shifts', 'what's the angle', 'turn this source into short-form ideas', 'what should this say', or hands over a source to repurpose for cold traffic. Do NOT use when: the user wants a FINISHED piece rendered in a specific format (→ that format's skill — carousel / reel), just a hook with no shift work (→ write-hook), or a full long-form argument structure (→ writing-logic).

precision-principle

Writing employee · Wording sharpener for ONE claim, promise, or headline at a time — the Meaning Test (would two people read it the same way?), Word Force (use the lowest-force word that does the job), and precision language rules (kill qualifiers, uncertainty markers, two-part negations). USE WHEN user says 'meaning test', 'word force', 'tighten this claim', 'make this specific', 'make this measurable', 'precision check', 'is this claim clear', 'this sounds hypey', 'lower the intensity', or when an offer/CTA/headline/promise needs its wording sharpened. Scope: ONE sentence at a time — hand anything draft-sized to the writing chain.

retrospective

Builder employee · Capture learnings from skill usage and propose SKILL.md improvements - the door for patching an OBSERVED misfire. USE WHEN user says 'retrospective', 'this skill messed up', 'fix this skill', 'heal this skill', 'the skill got it wrong', 'that run was wrong', 'capture what went wrong', 'skill-level post-mortem', or when a skill just produced wrong output and the user wants its instructions patched. Do NOT use for a deliberate change with no observed failure - 'update skill X', 'add a rule to X', 'improve this skill' route to meta-create-skill. Making a skill reliable going forward (gates, enforcement, hardening) is also meta-create-skill's job, not a retrospective.

roadmap

Builder employee · The AIEB build-roadmap and triage front door. ALWAYS USE for 'what's next', 'what should I build next', 'what now', 'where am I', 'roadmap', 'continue building', 'next step', 'what should I build now', 'look at what I have and tell me what to build', 'highest-value thing to build next', or 'I don't know what to build'. Ladder mode tracks progress after onboarding and shows one next move. Triage mode turns a plain-language pain into the right skill, system, harness, MCP, hook, or command and routes to its builder. Also use when the user says 'I keep doing this by hand', 'I keep doing X by hand', describes a weekly recurring manual task, or asks what to build for repeated manual work without naming the artifact. Do NOT use for first-run setup (onboard), an explicit operational business diagnosis (business-x-ray), or when the exact artifact is already named (use its meta-create builder).

seed-framework

Writing employee · Universal teaching framework (Show → Extend → Expose → Direct) for explaining any concept in visuals, text, slides, video, or docs. Three modes — CREATE, DIAGNOSE, REWRITE. USE WHEN user says 'explain this concept', 'explain this so a beginner gets it', 'explain it simply', 'teach this', 'break this down', 'make this clear', 'SEED this', 'apply SEED', 'how do I explain X', 'this needs to be clearer', 'check if this makes sense', 'diagnose this content', 'find the gaps', 'would a reader understand this', 'SEED diagnose', 'SEED check', or when any concept needs to be taught to an audience who doesn't already understand it.

setup-aieb

Connect this device to AI Employee Builder and start first-run workspace setup. USE WHEN the user says "set up AIEB", "connect AIEB", "activate AI Employee Builder", "finish AIEB setup", or has just installed the plugin.

setup-content-employees

Setup employee · First-run onboarding for the Community content employees. Sets up everything the writing fleet needs on this machine — voice samples, brand positioning doc, visual identity (brand colors + fonts for your diagrams and graphics), output folders, and a routing map in CLAUDE.md — through a short guided interview. USE WHEN user says 'set up my content employees', 'onboard my writing employees', 'set up my voice', 'calibrate my voice', 'my voice files', 'set my brand colors', 'customize how my diagrams look', 'my visual style', 'get started with the writing skills', 'the write skill says my voice files are missing', or right after installing/upgrading to the Community plan. Safe to re-run any time — it updates only what you approve and never overwrites your work. Also USE WHEN user says 'calibrate my voice', 'calibrate my voice files', 'set up my voice samples'.

whale-bait

Writing employee · The Whale Bait lens — turn a content idea or a perspective shift into content that attracts the few RIGHT buyers (whales) instead of a broad crowd (fish), and generate fresh whale-bait ideas from the brand. Diagnoses fish-vs-whale and value-vs-insight, then charges the idea with the brand's Plan/Code/Ideas/Proof. USE WHEN user says 'make this whale bait', 'whale-bait this', 'is this fish bait', 'will this attract whales', 'better leads not more leads', 'give me whale bait ideas', 'content ideas for buyers', 'this attracts the wrong people', 'make this attract buyers not browsers', or hands over an idea / perspective-shift card to sharpen for the right audience. Runs AFTER perspective-shift (charges the shift it scoped) and BEFORE the format skill (write / carousel / reel-creator). Do NOT use to: render a finished piece (→ that format's skill), find the shift itself (→ perspective-shift), tighten a single claim (→ precision-principle), or write a full argument (→ writing-logic).

write

Writing employee · Write newsletters, long-form articles, daily emails, broadcast emails, and general one-off emails. USE when the requested deliverable is a newsletter, article, long-form post, daily email, broadcast email, or general email draft. Scope is editorial prose from idea to finished draft. Commercial product and offer copy is outside scope. Formatting-only work on an existing draft is outside scope. Planning-only and voice-setup jobs are outside scope.

write-hook

Writing employee · Write hooks and short-form CTA openings. One gated procedure (the PSF Hook Scaffold) voiced through five accents — 5-Beat Claim, Empathy, Framework, Story, PADAI Offer — as either the opening of a longer piece OR a complete short-form CTA piece. USE WHEN user says 'write a hook', 'write the opening', 'hook for this', '5-beat hook', 'carousel hook', 'video hook', 'proof slide', 'evidence structure', 'opening for this piece', 'slide 2', 'grab attention', 'short-form CTA', 'PADAI', 'offer email opener', 'sales email opener', 'one-day open', 'launch email opener'. Do NOT use for: a full announcement post (→ write), building a carousel (→ carousel), reel ideation or a full reel (→ reel-creator), a full sales email/sequence or sales page (→ copywriter), or full argument structure (→ writing-logic).

writing-format

Writing employee · Format an already-written draft for a publishing platform. ALWAYS USE when the user says 'format this for Substack', 'format for Substack', 'format this newsletter', 'apply formatting', 'make this publishable', 'format carousel', 'carousel formatting', 'Substack formatting', or asks only for platform-specific headers, bold, italics, blockquotes, spacing, or visual rhythm. This skill formats existing prose; it does not create the draft.

writing-humanize

Writing employee · Make AI-generated text sound human. Scrub AI tells (em-dashes, staccato, filler-deficit, announcer scaffolding, And/So/Now/Because openers, billboard proclamations), enforce active voice, and check deliverable integrity. Works standalone on any text. USE WHEN user says 'humanize this', 'clean up AI-isms', 'make this sound human', 'voice cleanup', 'remove AI patterns', 'sound more natural', 'check for AI tells', 'fix the voice', or when any writing needs voice cleanup after drafting. Do NOT use to install the creator's personal voice - 'make it sound like me', 'install my voice' route to writing-voice (humanize scrubs AI tells; voice install is a separate pass).

writing-logic

Writing employee · Argument scaffolding system. Takes an idea (or a source) and sequences it into a Perspective-Shift scaffold — one protagonist's Old→New journey across labeled beats, one takeaway each, ordered to meet the reader where they are and break the idea down progressively. Build the scaffold, then QA it for logical integrity before any prose. USE WHEN user says 'plan argument', 'scaffold this', 'argument structure', 'plan the newsletter structure', 'sequence these ideas', 'logic check my writing', 'check my logic', 'expand this idea', 'map my ACTs', or when building the argument structure for any content piece before writing prose.

writing-voice

Writing employee · Install your teaching voice. PRIMARY TEST — every sentence must sound like someone speaking and pointing at things, not a billboard or press release. Then point at named things, run What → Why → Takeaway on every teaching moment, and replace every abstract noun with a specific named thing. The Spoken-and-Pointing test runs FIRST on every sentence — only after that does the Circle Test, Verb Test, and Teaching Triad apply. USE WHEN user says 'voice check', 'install voice', 'install teaching voice', 'make it sound like me', 'scan for vagueness', 'point at things', 'name the gap', 'pointing voice', 'less announcy', 'sounds like a billboard', 'too staccato', when writing prose from a writing-logic scaffold, or to install the teaching-on-a-board stance on existing prose. Also USE WHEN user says 'this reads vague', 'too vague - point at things'.