A Claude Code plugin that turns Claude into a Product Management Operating System. Manna Ray orchestrates 31 PM skills across 5 structured workflows, managing context files, workflow state, and skill chaining — so PMs can focus on decisions, not process.
Manna Ray wraps your product knowledge in five context files (product, company, personas, competitors, goals) and automatically injects the right context when you run PM skills. It tracks what you've done, chains skills into multi-step workflows, and nudges you when context files go stale.
You stay in Claude Code. Manna Ray handles the plumbing.
# Clone into your Claude Code plugins directory
git clone <repo-url> ~/.claude/plugins/manna-rayOpen Claude Code in your project directory and run:
/manna-init
This scaffolds:
context/— Five context files (product, company, personas, competitors, goals)outputs/— Organized output directories (discovery, strategy, specs, analytics, launch, productivity).manna-ray/state.json— Project state trackingCLAUDE.md— Configures Claude to work with Manna Ray
You'll be guided through populating your context files.
/manna-run prd-generator
Manna Ray validates that required context files exist, warns if any are stale, loads them automatically, runs the skill, saves the output, and records the run in your project history.
/manna-workflow start idea-to-sprint
Workflows chain skills together. Each step's output feeds into the next. Use /manna-workflow next to advance, /manna-workflow skip to skip a step, or /manna-workflow cancel to abandon.
/manna-status
Shows context file health, active workflow progress, and recent skill runs.
| Command | Description |
|---|---|
/manna-init |
Scaffold a new PM project with context files and state |
/manna-run [skill] |
Run a skill with automatic context injection |
/manna-run list [mode] |
List available skills, optionally filtered by mode |
/manna-workflow start [name] |
Begin a multi-step workflow |
/manna-workflow next |
Advance to the next workflow step |
/manna-workflow skip |
Skip the current step |
/manna-workflow cancel |
Cancel the active workflow |
/manna-workflow restart [name] |
Cancel and restart a workflow |
/manna-workflow status |
Show detailed workflow step progress |
/manna-workflow list |
List all available workflows |
/manna-context init |
Create context file templates |
/manna-context check |
Show context file health (missing, empty, stale) |
/manna-context update [file] |
Update a context file with new insights |
/manna-status |
Project dashboard — context, workflows, recent runs |
/manna-history [skill] |
Run history and output files, optionally filtered |
Take a vague market opportunity and turn it into a validated, testable solution.
user-interview-analyzer → research-synthesis-engine → jtbd-extractor
→ user-journey-mapper → opportunity-solution-tree
Align the company on what to build next quarter.
north-star-finder → landscape-mapper → competitive-profile-builder
→ tech-debt-evaluator → cogs-analyzer → quarterly-planning-template → roadmap-builder
Take an approved roadmap initiative to sprint-ready tickets.
prioritization-engine → prd-generator → ab-test-designer
→ technical-spec-writer → user-story-writer → launch-checklist-generator
Prerequisite: a roadmap output (outputs/strategy/roadmap-*.md) must exist.
Position and price a feature to win in the market.
swot-analysis-generator → positioning-statement-generator → pricing-strategy-analyzer
→ sales-enablement-kit → weekly-plan → daily-plan
Analyze post-launch data and feed learnings back into the system.
ab-test-analyzer → funnel-analyzer → metric-framework-builder
→ research-synthesis-engine → backlog-prioritizer
| Skill | What It Does |
|---|---|
user-interview-analyzer |
Transform interview transcripts into structured snapshots |
research-synthesis-engine |
Combine insights from multiple research sources |
jtbd-extractor |
Turn raw research into Jobs-to-be-Done statements |
user-journey-mapper |
Map end-to-end user journeys across lifecycle stages |
opportunity-solution-tree |
Map outcomes to opportunities to testable solutions |
| Skill | What It Does |
|---|---|
north-star-finder |
Identify the one metric that captures core product value |
landscape-mapper |
Map competitors into positioning matrices |
competitive-profile-builder |
Build strategic competitor profiles |
swot-analysis-generator |
Structured SWOT with strategic implications |
positioning-statement-generator |
Positioning via April Dunford's framework |
pricing-strategy-analyzer |
Competitive pricing, packaging, revenue modeling |
roadmap-builder |
Structured roadmap linked to business objectives |
quarterly-planning-template |
Quarterly plan with themes, bets, resource allocation |
backlog-prioritizer |
Prioritize backlogs into P0/P1/P2 using RICE |
prioritization-engine |
Defend roadmap with RICE/ICE scoring frameworks |
tech-debt-evaluator |
Quantify business cost of tech debt |
cogs-analyzer |
Evaluate unit economics and infrastructure costs |
| Skill | What It Does |
|---|---|
prd-generator |
Transform ideas into structured PRDs |
technical-spec-writer |
Architecture, data models, API designs |
user-story-writer |
Sprint-ready INVEST stories with acceptance criteria |
launch-checklist-generator |
Launch checklists customized to feature tier |
| Skill | What It Does |
|---|---|
ab-test-designer |
Design experiments with hypotheses and sample sizes |
ab-test-analyzer |
Interpret results with ship/no-ship recommendations |
funnel-analyzer |
Identify drop-off points and optimization opportunities |
metric-framework-builder |
Design metrics frameworks showing what to measure |
| Skill | What It Does |
|---|---|
daily-plan |
Time-blocked daily plan aligned to goals |
weekly-plan |
Weekly priorities with milestones and risk tracking |
meeting-notes-processor |
Extract action items and decisions from meetings |
executive-update-generator |
Concise executive updates using SCARF framework |
sales-enablement-kit |
Battlecards and how-to-sell materials |
stakeholder-simulator |
Feedback from simulated CTO, UX, Sales perspectives |
Manna Ray uses five canonical context files that accumulate product knowledge over time:
| File | Contains |
|---|---|
context/product.md |
Product overview, roadmap, key metrics, known issues, positioning |
context/company.md |
Strategic priorities, team structure, business model, constraints |
context/personas.md |
User archetypes, jobs-to-be-done, pain points, current solutions |
context/competitors.md |
Competitor profiles, win/loss themes, market gaps, pricing intel |
context/goals.md |
Annual goals, quarterly OKRs, success metrics, blockers |
Context files are checked for staleness (30-day threshold). When a skill produces insights relevant to a context file, Manna Ray suggests updating it.
After initialization, your project looks like this:
your-project/
├── .manna-ray/
│ └── state.json # Workflow state, context checksums, run history
├── context/
│ ├── product.md # Product context
│ ├── company.md # Company context
│ ├── personas.md # User personas
│ ├── competitors.md # Competitive intel
│ └── goals.md # Goals & OKRs
├── outputs/
│ ├── discovery/ # Interview snapshots, JTBD, research synthesis
│ ├── strategy/ # Roadmaps, positioning, competitive analysis
│ ├── specs/ # PRDs, tech specs, user stories
│ ├── analytics/ # A/B tests, funnels, metrics
│ ├── launch/ # Sales kits, launch checklists
│ └── productivity/ # Daily/weekly plans
├── research/ # Raw research materials
└── CLAUDE.md # Configures Claude for Manna Ray
Manna Ray is a native Claude Code plugin — no compilation, no build step. Everything is markdown, bash, JSON, or YAML.
manna-ray/
├── .claude-plugin/
│ └── plugin.json # Plugin manifest
├── commands/ # Slash commands (auto-discovered)
├── skills/ # PM skills (prompt-driven .md files)
├── agents/ # Subagent definitions
├── hooks/
│ └── hooks.json # SessionStart hook
├── scripts/ # Bash helpers for state, context, workflows
├── workflows/ # YAML pipeline definitions
├── templates/ # Scaffolding templates
└── tests/ # Bash test suite
jq— JSON manipulation in scriptsyq— YAML parsing for workflow definitions
bash tests/test-state.sh
bash tests/test-context.sh
bash tests/test-workflow.sh
bash tests/test-session-start.sh30 tests covering state CRUD, context validation, workflow engine, and session detection.
- Native plugin model — No compilation. Markdown + bash + JSON + YAML.
- Skills stay as prompts — The plugin handles plumbing; Claude handles reasoning.
- Declare, don't discover — Dependencies declared in frontmatter and workflow YAML.
- Suggest, don't force — Warnings and prompts are always skippable.
- Portable paths —
${CLAUDE_PLUGIN_ROOT}and${CLAUDE_PROJECT_DIR}everywhere. - Incremental population — Context files grow organically over time.
MIT