im-shashanks/claude-plugins / shaktra/skills/shaktra-pm
shaktra-pm
VerifiedProduct Manager workflow — PRD creation, user research analysis, persona generation, journey mapping, and feature prioritization. Entry point for product definition work.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 7.1 KB |
| agent-prompts.md | markdown | 3.4 KB |
| brainstorm-workflow.md | markdown | 4.9 KB |
| full-workflow-hypothesis.md | markdown | 5.0 KB |
| full-workflow-research.md | markdown | 4.8 KB |
| full-workflow.md | markdown | 3.0 KB |
| guided-entry.md | markdown | 5.5 KB |
| journey-workflow.md | markdown | 4.9 KB |
| persona-workflow.md | markdown | 4.0 KB |
| prd-workflow.md | markdown | 5.7 KB |
| prioritization-workflow.md | markdown | 5.6 KB |
| research-workflow.md | markdown | 5.2 KB |
| templates/prd-one-page.md | markdown | 1.7 KB |
| templates/prd-standard.md | markdown | 3.5 KB |
prd-workflow.md
View on GitHub# PRD Creation Workflow
Standalone workflow for creating Product Requirements Documents. Can be invoked directly or as Phase 2 of the full PM workflow.
## Overview
PRD creation follows a structured process: gather context, select template, guide through sections, quality review, and write. The output is a schema-compliant PRD ready for TPM consumption.
---
## Steps
### Step 1 — Load Context
Read:
- `.shaktra/settings.yml` — project type, language, architecture
- `.shaktra/pm/brainstorm.md` — if exists, use as input
- `.shaktra/memory/principles.yml` — prior principles
- `.shaktra/research/synthesis.yml` — if exists, for user insights
- `.shaktra/personas/*.yml` — if exist, for user references
### Step 2 — Template Selection
Determine which template to use:
**Use AskUserQuestion tool:**
Ask the user: "Which PRD template fits your feature?"
- **Standard PRD (6-8 weeks)** — Full PRD for complex features requiring design docs
- **One-Page PRD (2-4 weeks)** — Abbreviated PRD for smaller, well-understood features
Wait for user response before proceeding.
**Auto-selection heuristics (if user doesn't specify):**
- Greenfield project → Standard
- Enhancement to existing feature → One-Page
- Multiple external integrations → Standard
- Single-component change → One-Page
Templates are in `templates/prd-standard.md` and `templates/prd-one-page.md`.
### Step 3 — Guided Input
For each PRD section, gather information from the user. Use brainstorm notes as defaults where available.
**Section flow (Standard template):**
1. **Problem Statement**
- Use brainstorm `problem` section if available
- Ask for evidence points if not in brainstorm
2. **Users & Personas**
- Reference existing personas if they exist
- Otherwise, ask: "Who are the target users?"
3. **Goals & Success Metrics**
- Ask: "What does success look like? How will you measure it?"
- Ensure at least one quantifiable metric
4. **Functional Requirements**
- For each requirement, capture: description, MoSCoW priority, acceptance test
- Start with "Must Have" — what's blocking launch?
- Then "Should Have" — important but not blocking
- Then "Could Have" — nice to have
- Finally "Won't Have" — explicit exclusions
5. **Non-Functional Requirements**
- Ask about: performance, scalability, reliability, security
- Only include what's relevant — don't force all categories
6. **Scope**
- Explicitly state in-scope and out-of-scope
- Use brainstorm `opportunity.out_of_scope` if available
7. **Assumptions & Constraints**
- Technical, business, timeline, resource constraints
- Assumptions that, if wrong, would change the approach
8. **Risks & Mitigations**
- Identify top 3-5 risks
- For each: likelihood, impact, mitigation strategy
9. **Dependencies** (optional)
- External teams, systems, decisions needed
10. **Timeline** (optional)
- High-level milestones only — not sprint-level
### Step 4 — Generate PRD
Spawn **product-manager** (mode: prd-create) with:
- Gathered inputs
- Selected template
- Settings context
PM generates the PRD following the template structure.
### Step 5 — Quality Review
Run quality loop:
```
attempt = 0
max_attempts = 3
WHILE attempt < max_attempts:
attempt += 1
# Review against schema rules
findings = review_against_prd_schema()
IF no findings:
EMIT QUALITY_PASS
BREAK
# Fix findings
FOR each finding:
IF fixable by PM:
auto_fix(finding)
ELSE:
collect_user_input(finding)
CONTINUE
IF attempt == max_attempts AND findings remain:
EMIT MAX_LOOPS_REACHED
present_findings_to_user()
```
**Schema validation rules:**
| Rule | Severity |
|---|---|
| All "must" requirements have acceptance_test | P0 |
| At least one measurable success metric | P0 |
| Problem statement defines target user | P1 |
| Every requirement has unique ID | P1 |
| Scope has both in-scope and out-of-scope | P1 |
### Step 6 — Write PRD
Write the approved PRD to `.shaktra/prd.md`.
If a PRD already exists:
- Ask user: "PRD already exists. Replace or create new version?"
- If replace: overwrite
- If new version: increment version in frontmatter, archive old version
EMIT `PRD_COMPLETE`
### Step 7 — Present for Approval (Full Workflow Only)
If running as part of full workflow:
- Present PRD summary to user
- Wait for explicit approval
- EMIT `PRD_APPROVED` on approval
If standalone:
- Present PRD summary
- Recommend next step: `/shaktra:tpm` for design and stories
---
## Completion Report
```
## PRD Created
**File:** .shaktra/prd.md
**Template:** {standard | one-page}
**Status:** {draft | approved}
### Summary
- **Problem:** {1-line problem statement}
- **Users:** {primary user segment}
- **Requirements:** {count must} must, {count should} should, {count could} could
### Quality
- Review iterations: {count}
- Final status: PASS
### Validation Warnings
- {any P1 findings accepted by user}
### Next Steps
1. Review .shaktra/prd.md and refine as needed
2. Run `/shaktra:tpm` to create design docs and stories
```
---
## PRD Updates
To update an existing PRD:
1. User invokes `/shaktra:pm prd` with existing PRD
2. Detect existing PRD at `.shaktra/prd.md`
3. Ask: "Update existing PRD or create from scratch?"
4. If update:
- Load current PRD
- Ask which sections to modify
- Apply changes, increment version
- Re-run quality review
5. Write updated PRD
---
## Integration with TPM
The TPM workflow (`/shaktra:tpm`) consumes the PRD:
- Reads `.shaktra/prd.md` as input
- Architect uses PRD for design context
- PM uses PRD for gap answering and coverage analysis
- Stories reference PRD requirements via `REQ-XXX` IDs
Ensure requirements have stable IDs — changing IDs breaks story references.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| add-product-e2e-evalpaperclipai/paperclipAdd or extend a Paperclip full-stack runner E2E workflow, fixture, matcher, or report evidence path for local or Daytona execution. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| add-runner-evalpaperclipai/paperclipAdd or extend a Paperclip Runner protocol evaluation definition, roster, assertion, or report fixture with provenance and narrow validation. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| diagnose-why-work-stoppedpaperclipai/paperclipCompatibility entry point for failed, stalled, looping, or over-recovered Paperclip work. Use fix-failed-paperclip-run for the full diagnosis, regression, repair, real verification, and green-PR workflow; honor an explicitly requested diagnosis-only scope. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |
| fix-failed-paperclip-runpaperclipai/paperclipDiagnose and fix failed, stuck, slow, or incorrectly blocked Paperclip tasks and runs. Use with a task/run URL or failure report to trace the root cause, prove a regression red then green, check second-order effects, verify the real workflow, and deliver a PR with passing checks. | Stars: 115k | Updated | Claude CodeCodex | Verified | Install |