atlassian-pm

Claude Code plugin for AI-powered Jira & Confluence automation — create Epics, Tasks, and manage Scrumban flow using natural language.

🆕 What's New in v3.16.x

v3.16.0 — Dual-Zone Acceptance Criteria (2026-04-17) Every Jira ticket's AC now splits into two first-class H3 zones:

Business (มุมธุรกิจ/PM/ผู้ใช้) — observable outcomes only; IDs B-AC1..N
Developer (มุม dev/QA/AI agent) — testable specs with SLA/channels/patterns; IDs Dev-AC1..N with (derived from B-ACx) cross-refs

Applies to all create/update skills for Epic / Story / Task / Subtask / Bug. Ends business-vs-technical AC debate — PM gets clean zone, AI agent + dev + QA get unambiguous executable spec.

v3.16.1 — Hotfix: Grandfather for Existing Users (2026-04-17)

New validator check S7 (markdown-in-text) now defaults to WARN instead of ERROR — existing tickets with legacy text blobs won't suddenly fail
Opt into strict mode: AdfValidator(markdown_strict=True) or CLI --markdown-strict
S8 (dual-zone AC structure) also warn-only by default; CLI --dual-zone-strict
Defaults flip to ERROR in v3.18.0 — plan your migration window

See CHANGELOG.md for full details + migration notes.

💡 Idea & Concept

Vibe Coding for Jira — แปลง idea เป็น backlog ได้ใน 2 interactions โดยไม่ต้องเขียน Jira ticket เอง

💡 "สร้าง feature push notification บนมือถือ"
   → /vibe-full → Epic + Tasks + AI-Ready child Tasks
   → พร้อม Implementation Hints (entry point, pattern, test command)
   → ทีมงานแค่ /implement {{PROJECT_KEY}}-123 ใน Claude Code

Core Concepts:

Concept	สิ่งที่ทำ
Vibe Mode	Fast path: idea → backlog ใน 2 interactions (default สำหรับทุก creation skills)
AI-Ready Tasks	แต่ละ Task มี Implementation Hints — entry point file, pattern ที่ควรใช้, test command
Quality Gate	ADF ต้องผ่าน QG ≥ 90% ก่อน publish — ป้องกัน format พัง
Hard Rules	10 rules ที่ hooks enforce อัตโนมัติ — ป้องกัน silent failures
Token Cache	SQLite cache ลด Jira API calls 80–90% — ประหยัด tokens
Domain Expertise	แต่ละ skill มี Scrum/SAFe/ITIL/DORA/IEEE 829 notes ฝังอยู่

ใครควรใช้:

PM/PO ที่ต้องการแปลง idea → backlog ไว
Engineer ที่ต้องการ Tasks พร้อม implementation hints
QA ที่ต้องการ test plans จาก acceptance criteria
Team ที่ใช้ Scrumban และต้องการ flow management

Quick Start

/plugin marketplace add wasikarn/atlassian-pm
/plugin install atlassian-pm@atlassian-pm
/atlassian-pm:apm-setup

Claude will ask for your Jira site, project key, and board ID — then configure everything automatically.

After plugin updates: re-run /atlassian-pm:apm-setup to rebuild the atlassian-cache venv. The MCP server starts automatically — setup just installs its Python dependencies. Requires Claude Code with plugin support. If /plugin install is unavailable, see Manual Installation.

How It Works

⚡ Commands:  /vibe-full → chains skills end-to-end with confirmation gates
🚀 Vibe mode: /atlassian-pm:apm-vibe-plan "feature desc" → Epic + Tasks + AI-Ready child Tasks
   Skills:   /atlassian-pm:apm-create-task → Explore codebase → Write ADF → QG ≥ 90% → Jira

Describe in natural language (Thai/English) or use a Command for full workflow
Explore — Claude finds real implementation paths in your codebase
Quality Gate — ADF scored ≥ 90% before publish (hooks enforce 10 hard rules)
Publish — acli creates issues, MCP handles field updates

Workflow

flowchart TD
    A([💬 User Intent]) --> CMD["⚡ Commands Fast-Path\n/vibe-full · /story-full · /epic-full · /blueprint-full\n/bug-full · /sprint-plan-full · /sprint-close-full\n/release-full · /tech-debt-full · /story-analyze-full"]
    A --> B{New or Existing?}
    CMD -.->|"auto-chains skills below"| M
    VP --> V

    B -->|New| C["/search-issues\ndedup check"]
    B -->|Existing| D{Edit scope?}
    B -->|"Confluence spec"| SS["/spec-to-stories"]

    D -->|Single issue| E["/update-{epic,task}"]
    D -->|Need new child Tasks| VF["/verify-issue"]
    D -->|Task + children sync| F["/sync-artifacts"]

    E --> V["/verify-issue"]
    VF --> V
    F --> V
    SS --> V

    C --> G{Scope?}
    G -->|"Greenfield / Architecture"| H["/blueprint\nConfluence + backlog map"]
    G -->|"Unclear / High-risk"| I["/refine-epic\n4-role debate"]
    G -->|"Clear scope"| K["/create-task"]
    G -->|"Idea → tasks\none-shot"| VP["/vibe-plan 🚀"]
    G -->|"Bug report"| BT["/bug-triage"]
    G -->|"Task / Spike"| K

    H --> J["/create-epic"] --> K
    I --> K
    K --> L["/create-testplan\noptional"] --> V
    K --> V
    BT --> TP["/create-testplan"] --> V
    V --> M([✅ Jira + Confluence])

    subgraph flow["Scrumban Flow"]
        direction LR
        FC["/flow-check\n--replenish"] --> ST["/start-ticket"]
        ST --> SQ["/ship-to-qa"]
        SQ -.->|"sprint close"| Q["/close-sprint"] --> R["/retrospective-analyst"]
    end
    M -.->|"Pull when Ready queue low"| flow

    classDef skill fill:#dbeafe,stroke:#3b82f6,color:#1e3a5f
    classDef gate fill:#dcfce7,stroke:#16a34a,color:#14532d
    classDef endpoint fill:#f3f4f6,stroke:#6b7280,color:#111827
    classDef cmd fill:#fef9c3,stroke:#ca8a04,color:#713f12
    class C,E,F,VF,H,I,J,K,L,FC,ST,SQ,BT,TP,Q,R,SS,VP skill
    class V gate
    class A,M endpoint
    class CMD cmd

Commands

End-to-end orchestration chains — the fastest way to get things done. Each command chains multiple skills in sequence with confirmation gates between stages. Invoked as /name (no namespace prefix).

Command	Chains	Description
`/vibe-full`	search-issues → vibe-plan → verify-issue --with-subtasks	Idea → AI-Ready Tasks in one shot (vibe coding)
`/story-full`	search-issues → create-task → verify-issue --with-subtasks	Full Task creation with dedup + quality check
`/epic-full`	search-issues → create-epic → create-task → verify-issue --with-subtasks	Full Epic + Task creation end-to-end
`/blueprint-full`	blueprint → create-epic → create-task → verify-issue --with-subtasks	Greenfield feature from design to verified backlog
`/bug-full`	search-issues → bug-triage → create-testplan	Bug report with triage + test plan
`/qa-full`	create-testplan → execute-testplan	Create test plan + run against staging in one step
`/story-analyze-full`	verify-issue --with-subtasks	Verify existing Task + child alignment
`/sprint-close-full`	close-sprint → retrospective-analyst	Sprint closure + auto-generated retrospective
`/sprint-plan-full`	plan-sprint → map-dependencies	(Release forecasting only — not Scrumban daily flow) Capacity-based sprint allocation
`/release-full`	plan-release → release-notes	Release plan + Confluence release notes
`/tech-debt-full`	scan-tech-debt → create-task (per item)	Scan and create tasks for selected tech-debt items

Skills

Individual steps invoked as /atlassian-pm:<name>. Use when you need finer control over a specific phase. Each skill is a multi-phase workflow with domain-expert notes (Scrum, SAFe, ITIL, DORA, IEEE 829) embedded alongside the steps. Grouped by primary user role — many skills are useful across roles.

PM / Product Owner

Backlog ownership, flow management, documentation, and reporting.

Skill	Flags	Description
`/atlassian-pm:apm-blueprint`		5-role debate → Confluence blueprint + backlog map (S/M/L tiers)
`/atlassian-pm:apm-refine-epic`		4-role debate for unclear or high-risk requirements
`/atlassian-pm:apm-create-epic`	`--thorough` `--no-doc`	Epic + Confluence doc. Vibe default: auto-extract, skip RICE. `--no-doc` for Jira-only (skip Confluence). `--thorough` for full ceremony.
`/atlassian-pm:apm-vibe-plan`	`--dry-run`	🚀 Idea → Epic + Stories + AI-Ready Subtasks in one shot (max 2 interactions). `--dry-run` to preview plan without creating in Jira.
`/atlassian-pm:apm-plan-release`		Multi-sprint release plan + Confluence page + Jira Fix Version
`/atlassian-pm:apm-spec-to-stories 12345`		Convert Confluence spec page → batch-create Tasks
`/atlassian-pm:apm-search-issues`		Dedup check before creating
`/atlassian-pm:apm-assign-issue ABC-123 [name]`		Assign issue (bypasses MCP silent failure)
`/atlassian-pm:apm-flow-check`	`--replenish`	Board health snapshot + Scrumban replenishment trigger
`/atlassian-pm:apm-close-sprint`	`--sprint 123`	Close sprint: triage incomplete → move → Confluence review
`/atlassian-pm:apm-standup-report`	`--post`	Daily standup digest per assignee with anomaly detection
`/atlassian-pm:apm-reschedule-sprint`	`--sprint 123 --shift +7`	Bulk-shift issue dates across a sprint or issue list
`/atlassian-pm:apm-activity-report`	`--hours 48`	Work activity report from session history
`/atlassian-pm:apm-create-doc`		Create page: `tech-spec`, `adr`, `parent`
`/atlassian-pm:apm-update-doc`		Update or move a Confluence page
`/atlassian-pm:apm-release-notes`		Generate Confluence release notes from a Jira Fix Version
`/atlassian-pm:apm-plan-sprint`	`--sprint 123` `--carry-over-only`	(Release forecasting only) Capacity-based sprint allocation

Engineer / Tech Lead

Story and task authoring, DLC flow, codebase exploration, issue maintenance, and quality gates.

Skill	Flags	Description
`/atlassian-pm:apm-start-ticket ABC-123`	`--force`	Read AC + transition to In Progress. WIP gate enforced.
`/atlassian-pm:apm-ship-to-qa ABC-123`		Post PR + preview URLs to Jira + transition to Ready for QA. WIP gate enforced.
`/atlassian-pm:apm-create-task`	`--thorough`	Task with child Tasks. Vibe default: auto-extract + Implementation Hints. Auto-detects type (feature, bug, chore, spike).
`/atlassian-pm:apm-map-dependencies`	`--keys ABC-1,ABC-2`	Critical path + swim lane dependency analysis
`/atlassian-pm:apm-update-epic ABC-123`		Edit Epic — scope, RICE, metrics
`/atlassian-pm:apm-update-task ABC-123`		Edit Task — ACs, scope, description, format
`/atlassian-pm:apm-sync-artifacts ABC-123`		Bidirectional sync: Task ↔ child Tasks ↔ Confluence
`/atlassian-pm:apm-verify-issue ABC-123`	`--with-subtasks` `--fix` `--dry-run`	ADF format + INVEST criteria check
`/atlassian-pm:apm-scan-tech-debt`		Aggregate tech-debt/spike issues → Effort×Impact matrix on Confluence

QA / Tester

Test planning, bug intake, and acceptance verification.

Skill	Flags	Description
`/atlassian-pm:apm-create-testplan ABC-123`		Test Plan + `[QA]` Tasks from parent ACs
`/atlassian-pm:apm-execute-testplan ABC-123`		Run Google Sheet test cases via Playwright → write results back → create bug tickets
`/atlassian-pm:apm-bug-triage`	`--no-assign`	Full triage: intake → P1/P2/P3 severity → dedup check → assign. `--no-assign` to skip assignment gate.

Agents

Internal subagents dispatched automatically by skills and commands — not invoked directly. Organized in 3 layers by responsibility.

Agent	Model	Role
`code-explorer`	haiku	Codebase exploration; Memory-First Protocol
`issue-bootstrap`	haiku	Pre-fetch issue + parent + children context in one pass
`jira-search`	haiku	Duplicate detection with confidence scoring (EXACT/HIGH/MEDIUM/LOW)
`quality-gate`	sonnet	ADF quality scoring; Pattern Memory; Team Convention Check
`pr-description-writer`	haiku	Generate PR description from branch + issue
`pr-review-jira-sync`	haiku	Sync merged PR back to Jira (transition + comment)
`velocity-tracker`	haiku	Velocity history; anomaly detection (1.5σ); per-member stats
`sprint-transition-agent`	haiku	Batch sprint issue moves + sprint state transitions
`spec-parser-agent`	haiku	Parse Confluence spec → structured requirements (Read-only)
`story-writer`	sonnet	ADF JSON generation; Convention Memory; Service-Aware AC Defaults
`alignment-checker`	sonnet	AC Coverage Matrix; Predictive Risk Flags; Scope Drift detection
`backlog-groomer`	sonnet	WSJF scoring; aging alerts; Sprint-Ready/Blocked/Orphan grouping
`retrospective-analyst`	sonnet	Cross-Sprint Comparison; Team Health Score (0–100)
`sprint-planner`	sonnet	Risk-Adjusted Capacity; 3 Scenario Planning
`estimation-calibrator`	haiku	SP calibration from historical similarity; HIGH/MEDIUM/LOW confidence
`risk-forecaster`	sonnet	4-dimension delivery risk; named mitigations; adjusted scenarios
`adf-surgeon`	haiku	Structural ADF repair; 10 known Jira quirks; content-safe
`team-pattern-advisor`	sonnet	Multi-sprint strategic patterns; ≥3 data point threshold
`test-case-runner`	sonnet	Execute single Playwright test case; return structured result + evidence
`bug-evidence-writer`	haiku	Generate ADF bug description from test failure evidence

Usage Examples

Natural Language Prompts

Type these directly in Claude Code — no slash command needed. Claude detects intent and invokes the right skill automatically.

Feature Planning (vibe-plan)

vibe plan "coupon redemption at checkout for logged-in users"

สร้าง feature ระบบ push notification บนมือถือ

แตก feature video upload พร้อม progress bar ออกเป็น tasks ให้หน่อยครับ

plan feature: social login with Google OAuth — affects website + backend API

Each generates: Epic + Tasks + AI-Ready child Tasks with Implementation Hints (entry point, pattern, test command). Max 2 interactions — one tree review, then auto-creates in Jira.

Issue Verification (verify-issue)

verify {{PROJECT_KEY}}-123 --with-subtasks

ตรวจสอบ {{PROJECT_KEY}}-456 พร้อม child tasks ทั้งหมด

check alignment for {{PROJECT_KEY}}-789 --fix

technical quality check for {{PROJECT_KEY}}-321 — 3 services impacted: BE, Admin, Website

Checks ADF format, INVEST criteria, and alignment across the full issue tree. Use --fix to auto-repair issues found.

Task Creation

สร้าง task: ผู้ใช้สามารถ reset password ผ่าน email ได้

create task for Google SSO login — users should be able to sign in without a password

สร้าง task สำหรับ feature แสดง transaction history ใน admin dashboard [FE-Admin]

Bug Report

bug: ผู้ใช้ checkout ไม่ได้เมื่อ coupon code มีตัวอักษรพิเศษ

video player crash เมื่อ seek ไปที่ timestamp ที่ยังไม่ได้ buffer — P2

bug report: API returns 500 when user has more than 50 items in cart

Quality Check

verify {{PROJECT_KEY}}-123 --with-subtasks --fix

ตรวจสอบ ticket {{PROJECT_KEY}}-456 พร้อม subtasks ทั้งหมด และ fix ที่มีปัญหา

check if {{PROJECT_KEY}}-789 subtasks are aligned — dry run only

Slash Command Workflows

Full Feature (vibe path)

# Fastest: idea → AI-Ready subtasks in one shot
/vibe-full "real-time notification system for mobile users"
# → dedup check → Epic + Stories + subtasks → verified

Full Feature (thorough path)

# 1. Design (multi-role debate → Confluence + backlog map)
/atlassian-pm:apm-blueprint
→ "Build a real-time notification system"

# 2. Dedup check
/atlassian-pm:apm-search-issues

# 3. Epic + Confluence doc
/atlassian-pm:apm-create-epic

# 4. Task + child Tasks (explores codebase automatically)
/atlassian-pm:apm-create-task

# 5. QA child Tasks (optional)
/atlassian-pm:apm-create-testplan ABC-123

# 6. Verify the full tree
/atlassian-pm:apm-verify-issue ABC-123 --with-subtasks

Unclear Requirements

# 4-role debate before writing Jira artifacts
/atlassian-pm:apm-refine-epic
→ Roles: PO × Tech Lead × Engineer × QA
→ Output: revised requirements + refined ACs → ready for /create-task

Scrumban Flow

# Check board health + replenish Ready queue when low
/atlassian-pm:apm-flow-check

# Start a ticket (read AC + move to In Progress)
/atlassian-pm:apm-start-ticket {{PROJECT_KEY}}-123

# Ship to QA after PR is open (posts PR + preview URLs + transitions ticket)
/atlassian-pm:apm-ship-to-qa {{PROJECT_KEY}}-123

Update with Cascade

# Task only
/atlassian-pm:apm-update-task ABC-123

# Task + child Tasks + Confluence (all in sync)
/atlassian-pm:apm-sync-artifacts ABC-123

Online Installation (Recommended)

Install without cloning the repo — Claude handles everything.

Before you begin — have these ready:

Atlassian API token (Account Settings → Security → API tokens)

Jira site URL (e.g. your-company.atlassian.net)

Jira project key (e.g. ABC)

Board ID (optional — can look up after setup)

Setup takes 2–3 minutes (downloads acli, uv, and syncs the cache server venv). Claude Code must be restarted once after setup to activate the MCP server.

Claude Desktop (GUI)

Open Settings → Extensions → Add marketplace, enter:

wasikarn/atlassian-pm

Click Sync — Claude Desktop fetches the plugin catalog from GitHub. Then find atlassian-pm in the Extensions list and click Install.

Finally, open the Code tab and run setup:

/atlassian-pm:apm-setup

Claude Code (CLI)

Step 1 — Add marketplace

/plugin marketplace add wasikarn/atlassian-pm

Step 2 — Install plugin

/plugin install atlassian-pm@atlassian-pm

Step 3 — Run setup

/atlassian-pm:apm-setup

Claude will ask for your Jira site, project key, and board ID, then write the config and configure git filters automatically.

/atlassian-pm:apm-setup configures:

✓ acli (Jira CLI) — installed + authenticated
✓ mcp-atlassian — registered as user-scoped MCP server
✓ ~/.config/atlassian/.env — Jira/Confluence credentials
✓ git smudge/clean filters — placeholder conversion
✓ atlassian-cache venv — Python deps for local issue cache (SQLite + embeddings)

The atlassian-cache MCP server registers automatically via the plugin's .mcp.json — no manual step needed. Restart Claude Code once after setup to activate it.

Prerequisites

Tool	Purpose	Install
Homebrew	macOS package manager	`/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"`
Claude Code	AI agent runtime	`npm i -g @anthropic-ai/claude-code`
acli	Jira ADF publishing	`brew tap atlassian/homebrew-acli && brew install acli`
mcp-atlassian	Jira/Confluence MCP	configured by setup (Phase 4)
Python 3.11+	REST API scripts + cache server	`brew install python@3.11`

uv (Python package manager) is required for the cache server: brew install uv

Manual Installation

Click to expand (8 steps)

# 1. Clone
git clone https://github.com/wasikarn/atlassian-pm && cd atlassian-pm

# 2. Create config
cp config/project-config.json.template .claude/project-config.json
# Edit .claude/project-config.json with your Jira site, project key, board ID

# 3. Authenticate acli
echo "<api-token>" | acli jira auth login --site https://your-site.atlassian.net --email your@email.com --token

# 4. Create credentials file
mkdir -p ~/.config/atlassian && chmod 700 ~/.config/atlassian
cat > ~/.config/atlassian/.env << 'EOF'
JIRA_URL=https://your-site.atlassian.net
JIRA_USERNAME=your@email.com
JIRA_API_TOKEN=<your-api-token>
CONFLUENCE_URL=https://your-site.atlassian.net/wiki
CONFLUENCE_USERNAME=your@email.com
CONFLUENCE_API_TOKEN=<your-api-token>
EOF
chmod 600 ~/.config/atlassian/.env

# 5. Configure MCP
claude mcp add --scope user mcp-atlassian -- \
  uvx --no-cache mcp-atlassian==0.21.0 \
  --env-file ~/.config/atlassian/.env \
  --jira-projects-filter=YOUR_PROJECT_KEY \
  --confluence-spaces-filter=YOUR_SPACE_KEY

# 6. Run setup
./scripts/setup.sh

# 7. Install cache server venv
UV_PROJECT_ENVIRONMENT="$HOME/.claude/plugins/data/atlassian-pm-atlassian-pm/venv" \
  uv sync --project mcp-servers/atlassian-cache --extra embeddings

# 8. Load plugin
claude --plugin-dir /path/to/atlassian-pm
# Or add to ~/.claude/settings.json: { "pluginDirs": ["/path/to/atlassian-pm"] }

# Verify
acli jira auth status
/atlassian-pm:apm-doctor

Architecture

Coda-level performance via multi-tier state management:

L1 Cache: Process-level LRU cache for zero-latency reads.
State-Daemon: RAM-based state management via Unix Domain Sockets (Zero-Disk I/O).
SQLite WAL: ACID-compliant persistent storage with Write-Ahead Logging, Bloom Filters, and Delta Tracking.
Iterative ADF Processing: Stack-based traversal to prevent RecursionError in large documents.

Configuration

All values in .claude/project-config.json (gitignored). Template: config/project-config.json.template

File	Purpose
`.claude/project-config.json`	Jira fields, team roster, services, environments

Git Filter — automatic placeholder conversion:

Committed:    {{PROJECT_KEY}}-XXX   ← placeholders
Working tree: ABC-XXX   ← real values
Staged:       {{PROJECT_KEY}}-XXX   ← placeholders

Placeholder	Replaces
`{{PROJECT_KEY}}`	Your Jira project key (e.g., `ABC`)
`{{JIRA_SITE}}`	Your Jira domain (e.g., `acme-corp.atlassian.net`)

Project Structure

.claude-plugin/plugin.json       ← Plugin manifest
.mcp.json                        ← MCP server config
.claude/project-config.json      ← Real config (gitignored)

skills/                          ← 39 skills in 6 categories + 2 root-level
├── apm-setup/                   ← setup
├── apm-doctor/                  ← doctor
├── epic/                        ← blueprint, refine-epic, create-epic, vibe-plan, update-epic, plan-release
├── story/                       ← spec-to-stories, verify-issue, sync-artifacts
├── task/                        ← create-task, create-testplan, execute-testplan, bug-triage, start-ticket, ship-to-qa
├── sprint/                      ← flow-check, map-dependencies, close-sprint, standup-report, plan-sprint, retro-actions
├── confluence/                  ← create-doc, update-doc
└── utilities/                   ← search-issues, activity-report, scan-tech-debt, release-notes, status

references/                     ← Docs loaded on-demand (templates, hr-rules, troubleshooting)
scripts/                        ← Python scripts (ai/, api/, lib/, sprint/)
agents/                         ← 20 subagents (3-layer: Foundation, Analysis, Synthesis)
hooks/                          ← 68 Python hooks (HR1–HR10 enforcement, quality gates, cache)
.claude/commands/               ← 11 orchestration chains (vibe-full, story-full, epic-full, ...)
mcp-servers/atlassian-cache/    ← SQLite + FTS5 + embeddings cache

Full structure: See skills/README.md for skill index, hooks/README.md for hooks reference.

Tips

Always search first — run /atlassian-pm:apm-search-issues before creating anything to prevent duplicates.

Always verify after — run /atlassian-pm:apm-verify-issue ABC-XXX --with-subtasks to check ADF format, INVEST criteria, and alignment across the full tree.

Save tokens — run cache_sprint_issues(sprint_id=...) before flow-check or close-sprint to pre-cache all issues. Repeated reads cost 0 API tokens.

Let Claude explore — /atlassian-pm:apm-create-task always explores the codebase before creating child Tasks. Never skip — generic Tasks miss real implementation paths.

Dev hot-reload — after editing skill or agent files, use /reload-plugins in Claude Code.

Plugin development — marketplace.json must be in .claude-plugin/ (not repo root). When skills are organized in category subdirectories, skills must be an array of paths (e.g., ["./skills/setup/", "./skills/story/"]) — a single "./skills/" path will not discover nested categories.

Name		Name	Last commit message	Last commit date
Latest commit History 1,021 Commits
.claude-plugin		.claude-plugin
.claude		.claude
.github/workflows		.github/workflows
agents		agents
config		config
docs		docs
hooks		hooks
mcp-servers/atlassian-cache		mcp-servers/atlassian-cache
monitor		monitor
references		references
scripts		scripts
skills		skills
tests		tests
.cache		.cache
.gitattributes		.gitattributes
.gitignore		.gitignore
.markdownlint-cli2.jsonc		.markdownlint-cli2.jsonc
.markdownlint.json		.markdownlint.json
.markdownlintignore		.markdownlintignore
.mcp.json		.mcp.json
.next		.next
.shellcheckrc		.shellcheckrc
CHANGELOG.md		CHANGELOG.md
CLAUDE.md		CLAUDE.md
LICENSE		LICENSE
QUICKSTART.md		QUICKSTART.md
README.md		README.md
dist		dist
node_modules		node_modules
pyrightconfig.json		pyrightconfig.json
ruff.toml		ruff.toml

Folders and files

Latest commit

History

Repository files navigation

atlassian-pm

🆕 What's New in v3.16.x

💡 Idea & Concept

Quick Start

How It Works

Workflow

Commands

Skills

PM / Product Owner

Engineer / Tech Lead

QA / Tester

Agents

Usage Examples

Natural Language Prompts

Feature Planning (vibe-plan)

Issue Verification (verify-issue)

Task Creation

Bug Report

Quality Check

Slash Command Workflows

Full Feature (vibe path)

Full Feature (thorough path)

Unclear Requirements

Scrumban Flow

Update with Cascade

Online Installation (Recommended)

Claude Desktop (GUI)

Claude Code (CLI)

Prerequisites

Manual Installation

Architecture

Configuration

Project Structure

Tips

About

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases 100

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

Packages