setup-workflow
Scaffolds per-repo multi-agent collaboration framework + tool configuration — AGENTS.md, docs/adr/, PLAN.md, issue tracker, triage labels, domain docs.
Skill by atalariq
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/atalariq-setup-workflow
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: productivity, agent-building
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/setup-workflow
What's inside
- Skill name
setup-workflow- Description
- Scaffolds per-repo multi-agent collaboration framework + tool configuration — AGENTS.md, docs/adr/, PLAN.md, issue tracker, triage labels, domain docs. Run once per repo before using other planning/execution skills.
- License
- MIT
- Instructions
- 215 lines in SKILL.md · view on GitHub
Files (8)
- SKILL.md
- AGENTS_TEMPLATE.md · references/
- CLAUDE_TEMPLATE.md · references/
- domain.md · references/
- issue-tracker-github.md · references/
- issue-tracker-gitlab.md · references/
- issue-tracker-local.md · references/
- triage-labels.md · references/
SKILL.md (first lines)
# Setup Workflow — Multi-Agent Framework + Tool Config
Scaffold everything a repo needs so Hermes, OpenCode, and any other agent can collaborate:
- **Shared context** — AGENTS.md, CLAUDE.md, docs/adr/, PLAN.md, .hermes/
- **Git worktree convention** — isolated feature branches
- **Tool configuration** — issue tracker, triage labels, domain doc layout
This merges the old `setup-workflow` + `setup-skills` into one unified invocation. Run once per repo before using other skills.
## What This Creates
```
project/
├── AGENTS.md ← shared context for ALL agents
├── CLAUDE.md ← OpenCode adapter (references AGENTS.md)
├── docs/adr/
│ ├── index.md ← ADR log
│ └── (numbered ADRs) ← filled as decisions are made
├── PLAN.md ← active plan + progress tracker
├── .hermes/
│ ├── context.md ← Hermes session context
│ ├── blocker.md ← when stuck, write context here
│ └── session-notes/ ← per-session summaries
├── docs/agents/
│ ├── issue-tracker.md ← where issues live and how to use the CLI
│ ├── triage-labels.md ← five canonical role labels
│ └── domain.md ← domain doc consumer rules
└── .gitignore ← .worktrees/ + .hermes/session-notes/
```
## Workflow Pipeline
This skill is the entry point. After it runs, the pipeline is:
```
setup-workflow (done) → grill-plan → write-prd → write-todo → implement
```
## Process
### 1. Discover
Check what already exists. Don't assume anything:
- `git remote -v` — detect host (GitHub/GitLab) and infer repo
- `AGENTS.md` — does it exist? Read it.
- `CLAUDE.md` — does it exist? Read it.
- `docs/adr/` — does it exist? List contents.
- `PLAN.md` — does it exist? Read it.
- `.hermes/` — does it exist? List contents.
- `docs/agents/` — does this skill's prior output already exist?
- `.scratch/` — sign that a local-markdown issue tracker is in use
- `go.mod`, `package.json`, `Cargo.toml`, etc. — detect tech stack
- Is the project a git repo? (`git rev-parse --git-dir`)
- Is there at least one commit?
### 2. Present findings
Summarise what exists and what's missing. If AGENTS.md already has a `## Agent skills` block and all framework files exist, confirm with the user before making changes or ask if they want to reconfigure.
**If framework files already exist, skip Phase 1 and go straight to Phase 2.**
### 3. Phase 1 — Scaffold framework
For each missing file, create it using the templates in `references/`.
**CRITICAL:** Never overwrite an existing file without user confirmation. If a file exists, read it, show the user, and ask: "Replace, merge, or skip?"
Create these files:
**AGENTS.md** — from [AGENTS_TEMPLATE.md](./references/AGENTS_TEMPLATE.md). Fill in `{{PROJECT_NAME}}`, `{{STACK}}`, `{{REPO_URL}}`.
**CLAUDE.md** — from [CLAUDE_TEMPLATE.md](./references/CLAUDE_TEMPLATE.md). Fill in `{{PROJECT_NAME}}`, `{{STACK}}`.
**docs/adr/** — create directory with `index.md`:
```markdown
# Architecture Decision Records
| ADR | Title | Status |
|-----|-------|--------|
| 001 | [Title](001-title.md) | Proposed |
```
**PLAN.md** — create with initial structure:
```markdown
# Plan
## Current goal
[what we're working on]
## Progress
- [ ] Not started
## Notes
```
**.hermes/** — create with `context.md`, `.gitkeep` in `session-notes/`.
**Git worktree setup** — add to `.gitignore` if not present:
```bash
echo '.worktrees/' >> .gitignore
echo '.hermes/session-notes/' >> .gitignore
```
Explain the worktree convention:
```bash
git worktree add .worktrees/<feature-name> main
cd .worktrees/<feature-name>
# ... implement ...
git worktree remove .worktrees/<feature-name>
```
### 4. Phase 2 — Configure tools
Walk the user through three decisions **one at a time**. Present a section, get the answer, then move to the next.
#### Section A — Issue trackerFrom atalariq/skills, MIT.
More from atalariq/skills
- 0.502score
tdd
atalariq
Test-driven development with red-green-refactor cycles and vertical slices.
SkillCFreshness: C★ 1Claude Code +9 - 0.502score
improve-codebase
atalariq
Find deepening opportunities in a codebase — surface architectural friction and propose refactors that make it more testable and AI-navigable.
SkillCFreshness: C★ 1Claude Code +9 - 0.502score
implement
atalariq
Execute a validated issue or plan — write code against acceptance criteria, run tests continuously, cross-check ACs, self-review, and commit.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
grill-plan
atalariq
Stress-test your plan against the existing domain model — sharpens terminology, validates assumptions, updates CONTEXT.md and ADRs inline as decisions…
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
note-review
atalariq
Review one markdown note through retrieval practice — quiz the user from memory first, then reveal what their recall and the note itself are missing.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
code-review
atalariq
Review agent-generated code for spec compliance, test quality, architectural alignment, and agent anti-patterns.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
dispatch
atalariq
Dispatch multiple issues in parallel via subagents — each subagent runs implement on one issue independently.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
session-load
atalariq
Load a handoff document and resume context from a previous session.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
session-save
atalariq
Save the current session state to a handoff file so work can resume later.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
feynman
atalariq
Explain any concept in the simplest possible terms, then surface exactly where understanding is still fuzzy.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
flashcard-draft
atalariq
Turn a markdown note into an Anki-import-ready flashcard file — tab-separated text with Anki file headers, importable with zero hand-editing.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
worktree
atalariq
Manage git worktrees — create isolated feature branches, list active worktrees, clean up after merge.
SkillCFreshness: C★ 1Claude Code +9
Signals
- Stars
- 1
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Jun 10, 2026 (4 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/atalariq/skills
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from atalariq
- 0.502score
implement
atalariq
Execute a validated issue or plan — write code against acceptance criteria, run tests continuously, cross-check ACs, self-review, and commit.
SkillCFreshness: C★ 1Claude Code +9 - 0.502score
improve-codebase
atalariq
Find deepening opportunities in a codebase — surface architectural friction and propose refactors that make it more testable and AI-navigable.
SkillCFreshness: C★ 1Claude Code +9 - 0.502score
tdd
atalariq
Test-driven development with red-green-refactor cycles and vertical slices.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
grill-plan
atalariq
Stress-test your plan against the existing domain model — sharpens terminology, validates assumptions, updates CONTEXT.md and ADRs inline as decisions…
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
dispatch
atalariq
Dispatch multiple issues in parallel via subagents — each subagent runs implement on one issue independently.
SkillCFreshness: C★ 1Claude Code +9 - 0.497score
code-review
atalariq
Review agent-generated code for spec compliance, test quality, architectural alignment, and agent anti-patterns.
SkillCFreshness: C★ 1Claude Code +9
Cite this page
setup-workflow skill. Agent Mart. https://agentmart.dev/skills/atalariq-setup-workflow (last updated 2026-10-24)