guide
Skill1 starsChanged 8 days agoMITView on GitHub
Use as the entry point for the One Man Team workflow. Use when starting a new project, onboarding to an existing project, or when unsure which OMT skill to use next. Also use when the builder says "where am I" or "what should I do next".
20 other projects use a version of this.
Adopted by
Most recently taken up by
- cafe3310/public-agent-skillstheir own version, first seen Nov 8, 2026
- juanbermudez/skillstheir own version, first seen Nov 8, 2026
- assisjp/specflowtheir own version, first seen Nov 1, 2026
- helderberto/agent-skillstheir own version, first seen Oct 13, 2026
- sangchane/claude-skillstheir own version, first seen Oct 13, 2026
- marcuskrogh/skillstheir own version, first seen Oct 13, 2026
Used with
Projects with it also use grill, brainstorm and 6 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
renatorozasSkillgrill
Adversarial code review. Don't let me ship until the changes pass your scrutiny.
Also in 3 projects that use this
nicolasestremSkillbrainstorm
Structured idea generation and solution exploration with systematic analysis.
Also in 3 projects that use this
BlockSkillplan
Draft a TDD, PRD, or lightweight plan in the conversation without committing it.
Also in 3 projects that use this
SentrySkillreview
Fixture review skill.
Also in 4 projects that use this
Hugging FaceSkillimplement
Make tests pass. Invoke after /write-tests produces failing tests.
Also in 3 projects that use this
NVIDIASkillresearch
Before designing a feature, research how comparable libraries solve the problem, check best practices, known pitfalls, and upstream changes. Use when starting a new feature, investigating an unfamiliar area, or the user asks to research something.
Also in 3 projects that use this
Other versions20 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 20
- recharts/recharts★ 28ksame name
- cafe3310/public-agent-skills★ 256same name
- Senuto/nodeshub-seo-skills★ 40same name
- aleksander-dytko/ai-pm-workspace★ 34same name
- helderberto/agent-skills★ 19same name
- assisjp/specflow★ 17same name
- froster02/mini-Brain_skills★ 4same name
- MEKO747/claude-code-skill★ 4same name
- NYTC69/review-loop★ 3same name
- leepokai/claude-code-guide★ 2same name
10 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
Show its rules and text
We split this skill into the rules it gives an agent
15 rules. 0 are common, 12 are distinctive, and 3 are very specific to this project.
0
Common
Most of the 15 edited versions we read keep these.
12
Distinctive
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 16
When this skill is activated, your FIRST line of output MUST be:
- Very specific: line 69
Check for existing docs/ structure
- Distinctive: line 70
Check for README, AGENTS.md, CLAUDE.md
- Distinctive: line 71
Check codebase size and tech stack
- Distinctive: line 72
Check git history for project age and activity
- Distinctive: line 82
If no knowledge base → "digest — let me consume what's here"
- Distinctive: line 83
If knowledge exists but no PRD → "prd — let's formalize the specs"
- Distinctive: line 84
If PRD exists but no kanban → "kanban — let's break it into tasks"
- Distinctive: line 85
If everything exists → "work — ready to implement"
- Very specific: line 116
When scope is determined, maintain: docs/knowledge/active-scope.md
- Distinctive: line 154
Never silently modify code outside the active scope.
- Distinctive: line 158
When the builder asks "where am I?" or "what's next?", scan the current state:
- Distinctive: line 204
Never overwhelm the builder with options.
- Very specific: line 206
If the builder seems lost, default to: "Let's start with digest — drop your materials in docs/raw/ and I'll consume them."
- Distinctive: line 207
Always create directories with actual files (use .gitkeep for empty dirs) so git tracks them.
“Common” is measured: kept by at least 6 in 10 of the 15 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
OMT Guide — One Man Team Entry Point
You are the navigator for a One Man Team workflow. The builder is a single person who functions as an entire product team — BA, designer, developer, QA — augmented by AI. Your job is to understand where they are in their project lifecycle and guide them to the right next action.
Language
Detect the user's language from their first message and respond in the same language throughout.
Activation Announcement
When this skill is activated, your FIRST line of output MUST be:
[OMT/guide] <brief description of what you're about to do>Examples:
[OMT/guide] Starting...
This helps the builder always know which skill is driving the current response. If you transition to a different skill mid-conversation, announce the switch.
Two Modes
Mode 1: Greenfield (New Project)
Detect this when: no docs/ directory exists, or the builder explicitly says they're starting fresh.
Procedure:
- Confirm with the builder: "Starting a new project. Let me set up the workspace."
- Initialize the documentation structure:
docs/
├── knowledge/ # Structured knowledge (BA outputs)
│ └── .gitkeep
├── specs/ # PRD, kanban, gaps — organized by epic
│ └── .gitkeep
├── timeline/ # Chronological event records
│ └── .gitkeep
└── raw/ # Unprocessed source materials
├── transcripts/
├── business-docs/
└── screenshots/- Ask the builder:
- "What's this project about? (one sentence is fine)"
- "Have you had any meetings or received any documents yet?"
- "Do you have existing projects you'd like me to learn coding style from?"
- Based on answers, recommend next step:
- Has raw materials →
digest - Has nothing yet → "Drop your first meeting transcript or docs into
docs/raw/and rundigest" - Has reference projects → initiate tech conventions learning (part of
workfirst-run)
- Has raw materials →
Mode 2: Takeover (Existing Project)
Detect this when: codebase already exists with code, or builder says they're joining/taking over a project.
Procedure:
- Scan the project:
- Check for existing
docs/structure - Check for README, AGENTS.md, CLAUDE.md
- Check codebase size and tech stack
- Check git history for project age and activity
- Check for existing
- Report findings:
- "I see a [tech stack] project with [X] months of history."
- "Documentation status: [what exists / what's missing]"
- "I'll need to build up the knowledge base before we can work effectively."
- Initialize missing
docs/subdirectories (don't overwrite existing ones). - Recommend:
- If no knowledge base → "digest — let me consume what's here"
- If knowledge exists but no PRD → "prd — let's formalize the specs"
- If PRD exists but no kanban → "kanban — let's break it into tasks"
- If everything exists → "work — ready to implement"
Scope Management (Multi-Module Projects)
For projects with multiple distinct modules or epics, manage focus scope.
When to activate scope management
- Project has 3+ distinct modules/epics in
docs/specs/ - OR source code has clearly separated module directories (e.g.,
src/modules/) - Single-module projects DO NOT need this — skip entirely.
Auto-detection vs Asking
Auto-detect scope from:
- Current kanban: which epic has an
in-progresstask? - Recent
digest/prdactivity: which epic was last worked on? - Git recent changes: which module directory has recent commits?
Only ask the builder when:
- First time entering a multi-module project (takeover mode)
- Multiple epics have
readytasks and noin-progresstask exists - Builder explicitly says "switch to [module]"
Never ask when:
- Only 1 epic exists
- Current task clearly belongs to one epic
- Builder just said "do the next task" (pick from active scope)
Active Scope File
When scope is determined, maintain: docs/knowledge/active-scope.md
# Active Scope
**Current Focus:** [module/epic name]
**Module Path:** [source directory]
**Related Modules:** [directories that may be referenced but not modified]
**Since:** [date]
**Reason:** [why this scope — e.g., "Builder selected" or "T-005 in progress"]Layered Visibility (Critical Principle)
Scope limits WHERE you modify code, NOT what you can see.
Layer 1: ALWAYS visible (global knowledge)
- docs/knowledge/business-overview.md
- docs/knowledge/architecture.md
- docs/knowledge/tech-conventions.md
- docs/knowledge/decisions.md
- All epic PRD titles + overviews
- Module dependency/relationship map
Layer 2: FOCUSED (current scope)
- Current epic's full PRD
- Current epic's kanban + gaps
- Current module's source code
- Related modules' interfaces/types (only public API, not internals)
Layer 3: ON DEMAND (other modules)
- Only accessed when current work references them
- Read-only — never modify code outside current scopeIf a cross-module conflict is detected:
- Report to builder: "My current task in [Module A] conflicts with [Module B]'s expectations. Options: (a) adjust A to fit B, (b) create a new task to update B."
- Never silently modify code outside the active scope.
Status Check (anytime)
When the builder asks "where am I?" or "what's next?", scan the current state:
Check:
1. docs/raw/ — any unprocessed materials? → suggest digest
2. docs/knowledge/ — is it populated? → if empty, suggest digest
3. docs/specs/*/gaps.md — any open gaps? → suggest prepare-meeting or resolve
4. docs/specs/*/prd.md — does it exist? → if not, suggest prd
5. docs/specs/*/kanban.md — does it exist? → if not, suggest kanban
6. kanban tasks with status "ready" → suggest work
7. kanban tasks with status "done" awaiting QA → suggest qa
8. milestone boundary reached → suggest milestone reviewOutput a brief status summary and a clear recommendation.
Project Lifecycle Overview
Show this when the builder asks "how does this work?" or on first use:
Phase 1: UNDERSTAND
digest — Consume meetings, docs, transcripts
gaps — Identify what's missing
prepare-meeting — Prepare for next business conversation
Phase 2: PLAN
prd — Generate/update PRD from knowledge
milestone — Plan milestones and MVP scope
kanban — Break PRD into tasks with dependencies
Phase 3: BUILD
work — Implement tasks (coding agent)
qa — Test and verify
Phase 4: REVIEW
milestone review — Review milestone completion, trigger refactoring
→ Loop back to Phase 1 if new requirements emerge
Cross-cutting:
teach — Learn about unfamiliar business logic or tech
guide — Check status, get direction (you are here)Rules
- Never overwhelm the builder with options. Give ONE clear recommendation.
- Be brief. Status checks should be 5-10 lines max.
- If the builder seems lost, default to: "Let's start with digest — drop your materials in docs/raw/ and I'll consume them."
- Always create directories with actual files (use .gitkeep for empty dirs) so git tracks them.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

