skillstore-cli 1.0.0

package/data/free-skills/devflow-docs/plugin/skills/pdf-processor/assets/sample-output.md

@@ -0,0 +1,114 @@
+# Sample Output: Government IT Modernization RFP
+
+**Source**: `GOV-IT-Modernization-RFP-2026.pdf` (42 pages)
+**Extraction Quality**: HIGH (native text PDF)
+**Processed**: 42/42 pages
+
+---
+
+## Document Metadata
+
+| Field | Value |
+|-------|-------|
+| Title | Request for Proposal: IT Infrastructure Modernization |
+| Author | Department of Digital Services |
+| Created | 2026-01-15 |
+| Version | 2.1 (Final) |
+| Pages | 42 |
+
+---
+
+## Executive Summary (Pages 1-3)
+
+The Department of Digital Services seeks proposals for a comprehensive IT infrastructure modernization program. The program covers migration of 12 legacy systems to cloud-native architecture, implementation of zero-trust security framework, and establishment of a modern DevOps pipeline.
+
+**Budget Range**: $8.5M - $12M over 3 years
+**Submission Deadline**: 2026-04-30
+**Contract Duration**: 36 months with 2 optional 12-month extensions
+
+---
+
+## Extracted Requirements Table (Pages 8-15)
+
+| Req ID | Category | Description | Priority | Compliance |
+|--------|----------|-------------|----------|------------|
+| REQ-001 | Infrastructure | Migrate all 12 legacy systems to AWS GovCloud | Must Have | Mandatory |
+| REQ-002 | Infrastructure | Achieve 99.95% uptime SLA for Tier 1 systems | Must Have | Mandatory |
+| REQ-003 | Security | Implement zero-trust architecture per NIST 800-207 | Must Have | Mandatory |
+| REQ-004 | Security | FedRAMP High authorization for all cloud services | Must Have | Mandatory |
+| REQ-005 | DevOps | CI/CD pipeline with automated security scanning | Should Have | Scored |
+| REQ-006 | DevOps | Infrastructure as Code for all deployments | Should Have | Scored |
+| REQ-007 | Data | Data lake implementation for analytics workloads | Nice to Have | Scored |
+| REQ-008 | Data | Real-time dashboard for system health monitoring | Should Have | Scored |
+| REQ-009 | Training | Staff training program (min 200 personnel) | Must Have | Mandatory |
+| REQ-010 | Support | 24/7 Tier 3 support during transition period | Must Have | Mandatory |
+
+**Total Requirements Extracted**: 47 (10 shown above)
+- Must Have: 23
+- Should Have: 16
+- Nice to Have: 8
+
+---
+
+## Budget Section (Pages 18-22)
+
+### Cost Breakdown by Phase
+
+| Phase | Duration | Budget Allocation |
+|-------|----------|-------------------|
+| Phase 1: Assessment & Planning | Months 1-4 | $1.2M |
+| Phase 2: Infrastructure Build | Months 5-14 | $4.5M |
+| Phase 3: Migration & Testing | Months 15-26 | $3.8M |
+| Phase 4: Optimization & Handover | Months 27-36 | $2.5M |
+| **Total** | **36 months** | **$12.0M (max)** |
+
+### Budget Constraints
+- Travel expenses capped at 5% of total contract value
+- Hardware costs must not exceed 30% of Phase 2 budget
+- Training budget: minimum $500K allocated
+
+---
+
+## Timeline (Pages 23-27)
+
+### Key Milestones
+
+| Milestone | Target Date | Dependencies |
+|-----------|-------------|--------------|
+| Contract Award | 2026-06-01 | — |
+| Kickoff Meeting | 2026-06-15 | Contract signed |
+| Assessment Complete | 2026-09-30 | Kickoff |
+| Cloud Environment Ready | 2027-01-31 | Assessment |
+| First System Migrated | 2027-04-30 | Cloud ready |
+| 50% Systems Migrated | 2027-09-30 | First migration |
+| All Systems Migrated | 2028-03-31 | Progressive migration |
+| Performance Optimization | 2028-08-31 | All migrated |
+| Final Handover | 2028-12-31 | Optimization complete |
+
+---
+
+## Evaluation Criteria (Pages 30-35)
+
+| Criteria | Weight | Description |
+|----------|--------|-------------|
+| Technical Approach | 35% | Architecture design, migration strategy, security framework |
+| Past Performance | 25% | Similar projects, references, case studies |
+| Management Approach | 20% | Team structure, risk management, communication plan |
+| Cost Proposal | 15% | Total cost, cost realism, value for money |
+| Small Business Participation | 5% | Subcontracting plan, mentor-protégé arrangements |
+
+### Scoring Scale
+- Outstanding (5): Exceeds requirements, innovative approach
+- Good (4): Meets all requirements with minor strengths
+- Acceptable (3): Meets minimum requirements
+- Marginal (2): Fails to meet some requirements
+- Unacceptable (1): Fails to meet critical requirements
+
+---
+
+## Extraction Notes
+
+- All 42 pages processed successfully
+- 4 tables required multi-page merge (pages 8-15 requirements table)
+- 2 architecture diagrams detected on pages 28-29 (described inline)
+- Appendices A-D contain boilerplate terms (extracted but not shown in summary)

package/data/free-skills/devflow-docs/plugin/skills/pdf-processor/references/extraction-techniques.md

@@ -0,0 +1,115 @@
+# PDF Extraction Techniques
+
+## Text Extraction Strategies
+
+### Native Text vs OCR
+- **Native text PDFs**: Text is embedded as selectable characters. Direct extraction is fast and accurate.
+- **Scanned PDFs**: Pages are images. Requires OCR (Optical Character Recognition) for text extraction.
+- **Hybrid PDFs**: Mix of native text and scanned pages. Detect per-page and apply appropriate method.
+
+### Handling Multi-Column Layouts
+- Detect column boundaries by analyzing text block positions
+- Read columns left-to-right, top-to-bottom within each column
+- Watch for content that spans columns (titles, images, tables)
+- Use spatial analysis to determine reading order when columns are ambiguous
+
+### Text Cleaning
+- Remove headers/footers that repeat across pages
+- Handle hyphenation at line breaks (re-join split words)
+- Normalize whitespace and line breaks
+- Preserve paragraph boundaries
+
+## Table Extraction
+
+### Detecting Table Boundaries
+- Look for grid lines (ruled tables)
+- For borderless tables: detect alignment patterns in text blocks
+- Use whitespace gaps to identify column separators
+- Check for consistent row height patterns
+
+### Merging Cells and Spanning
+- Detect merged cells by analyzing grid structure
+- Handle column spans (header cells spanning multiple data columns)
+- Handle row spans (category labels spanning multiple rows)
+- Reconstruct logical table structure from visual layout
+
+### Complex Table Patterns
+- Nested tables (tables within cells)
+- Tables split across pages — detect continuation and merge
+- Tables with mixed text/numeric data
+- Tables with embedded lists or sub-items
+
+## Form Field Extraction
+
+### Field Types
+- **Text fields**: Extract label + value pairs
+- **Checkboxes**: Detect checked/unchecked state
+- **Radio buttons**: Identify selected option within group
+- **Dropdowns**: Extract selected value and available options if accessible
+- **Signature fields**: Flag presence, note if signed
+
+### Form Mapping
+- Associate labels with their corresponding fields
+- Handle multi-page forms with consistent field naming
+- Extract form metadata (form title, version, required fields)
+- Output as structured key-value pairs
+
+## Image/Diagram Extraction
+
+### Image Handling
+- Extract embedded images with position context
+- Determine image type (photo, diagram, chart, logo)
+- Provide descriptive caption based on surrounding text
+- Note image dimensions and quality
+
+### Diagram Interpretation
+- Architecture diagrams: identify components and connections
+- Flowcharts: extract process steps and decision points
+- Org charts: extract hierarchy and roles
+- Network diagrams: identify nodes and relationships
+
+## Metadata Extraction
+- **Author**: Document creator and last modifier
+- **Dates**: Creation date, modification date, print date
+- **Version**: Document version if specified
+- **Title/Subject**: From document properties
+- **Keywords/Tags**: If present in metadata
+- **Page count**: Total pages
+- **File size**: For reference
+- **PDF version**: Compatibility information
+- **Security settings**: Encryption, permissions (print, copy, edit)
+
+## Multi-Language Document Handling
+- Detect primary language from text content
+- Identify language switches within document
+- Apply language-specific OCR models for scanned content
+- Preserve original language in extraction, note detected language
+- Handle right-to-left (RTL) languages appropriately
+
+## Handling Scanned vs Digital PDFs
+
+### Detection
+- Check if text layer exists (try selecting text)
+- Analyze page content streams for text vs image objects
+- Test extraction: if native text yields garbage or nothing, likely scanned
+
+### Scanned PDF Best Practices
+- Pre-process images: deskew, denoise, enhance contrast
+- Use appropriate OCR engine and language model
+- Validate OCR output against expected patterns
+- Report confidence scores per page/section
+- Flag low-confidence sections for human review
+
+## Quality Assessment of Extraction Results
+
+### Metrics
+- **Completeness**: Percentage of pages/sections successfully extracted
+- **Accuracy**: Confidence score for OCR-based content
+- **Structure preservation**: Are headings, lists, tables intact?
+- **Data integrity**: Numbers, dates, names correctly extracted
+
+### Quality Flags
+- `HIGH`: Native text, clean extraction, structure preserved
+- `MEDIUM`: OCR with high confidence, minor formatting issues
+- `LOW`: OCR with low confidence, significant structure loss
+- `FAILED`: Unable to extract meaningful content from section

package/data/free-skills/devflow-docs/plugin/skills/pdf-processor/references/ocr-strategies.md

@@ -0,0 +1,167 @@
+# OCR Strategies Reference
+
+## OCR Engine Comparison
+
+### Tesseract OCR (Open Source)
+- **Best for**: High-quality scanned text, single-language documents, budget-conscious projects
+- **Accuracy**: 85-95% on clean scans, drops significantly on low-quality images
+- **Languages**: 100+ languages supported via trained models
+- **Cost**: Free
+- **Latency**: Moderate (1-5 seconds per page depending on complexity)
+- **Deployment**: Self-hosted, runs locally or in container
+- **Limitations**: Poor on handwriting, complex layouts, low-resolution images
+
+### Google Cloud Vision OCR
+- **Best for**: Mixed content (text + images), handwriting, complex layouts
+- **Accuracy**: 95-99% on most document types
+- **Languages**: 100+ languages with automatic detection
+- **Cost**: $1.50 per 1,000 pages (first 1,000/month free)
+- **Latency**: 1-3 seconds per page (network dependent)
+- **Key feature**: Document AI for structured extraction (invoices, receipts)
+
+### AWS Textract
+- **Best for**: Forms, tables, structured documents (invoices, tax forms)
+- **Accuracy**: 95-99%, especially strong on tables and forms
+- **Languages**: English primary, expanding language support
+- **Cost**: $1.50 per 1,000 pages (text), $15 per 1,000 pages (tables/forms)
+- **Latency**: 1-5 seconds per page
+- **Key feature**: Automatic table structure detection, key-value pair extraction
+
+### Azure Form Recognizer (Document Intelligence)
+- **Best for**: Pre-built models for common document types (invoices, receipts, ID cards)
+- **Accuracy**: 95-99% with pre-built models, customizable models improve further
+- **Languages**: 100+ languages for read, fewer for specialized models
+- **Cost**: $1.50 per 1,000 pages (read), $10 per 1,000 pages (pre-built models)
+- **Key feature**: Custom model training with as few as 5 labeled samples
+
+### Comparison Matrix
+| Feature              | Tesseract        | Google Vision    | AWS Textract     | Azure Form Recognizer |
+|----------------------|------------------|------------------|------------------|----------------------|
+| Table detection      | Poor             | Good             | Excellent        | Excellent            |
+| Form extraction      | None             | Basic            | Excellent        | Excellent            |
+| Handwriting          | Poor             | Good             | Good             | Good                 |
+| Custom training      | Yes (complex)    | AutoML           | Custom Queries   | Custom Models        |
+| On-premise           | Yes              | No               | No               | Yes (containers)     |
+| Batch processing     | Manual           | Batch API        | Async API        | Batch API            |
+
+## Table Detection
+
+### Algorithms and Approaches
+1. **Rule-based detection**: Look for horizontal/vertical lines forming grid patterns
+2. **ML-based detection**: Use object detection models (DETR, TableNet) to identify table regions
+3. **Hybrid**: Combine line detection with ML for robust results
+
+### Bordered Tables
+- Most reliable to detect: look for continuous horizontal and vertical lines
+- Use Hough Line Transform for line detection
+- Intersecting lines define cell boundaries
+- Process: detect lines -> find intersections -> define cells -> extract text per cell
+
+### Borderless Tables
+- Significantly harder: rely on text alignment and spacing
+- Use column detection via text x-coordinate clustering
+- Row detection via y-coordinate gaps
+- Whitespace analysis to determine column boundaries
+- Consider using ML-based approaches for higher accuracy
+
+### Multi-Page Tables
+- Detect table continuation markers ("continued on next page", repeated headers)
+- Track column positions across pages for alignment
+- Merge strategy: match columns by x-position, append rows
+- Watch for: header repetition, page numbers in table, footnotes breaking table
+
+## Form Field Extraction
+
+### Key-Value Pair Detection
+- **Label proximity**: Find text labels near input fields (left of, above)
+- **Colon detection**: "Name: John Doe" pattern - split on colon
+- **Spatial relationship**: Labels and values typically aligned horizontally or vertically
+- **Field grouping**: Related fields often share visual boundaries (boxes, sections)
+
+### Checkbox Detection
+- Look for small square regions (~10-20px) with or without fill
+- Filled checkbox: high pixel density inside square boundary
+- Empty checkbox: low pixel density, only border pixels
+- Common symbols: checkmark, X, filled square vs empty square
+
+### Signature Areas
+- Typically found at bottom of documents
+- Identified by: horizontal line above text "Signature", "Sign here"
+- Contain irregular strokes (unlike printed text)
+- May be blank (unsigned) or contain image data (signed)
+- Do not attempt OCR on signature areas - flag for human review
+
+## Multi-Language Document Handling
+
+### Language Detection Strategy
+1. **Pre-scan sampling**: OCR a small region, run language detection on output
+2. **Script detection**: Identify character scripts (Latin, CJK, Arabic, Devanagari)
+3. **Mixed documents**: Split into regions, detect language per region
+4. **Fallback**: Use multi-language model if detection is uncertain
+
+### Mixed-Language Documents
+- Common patterns: English headers with local-language body text
+- Process each text region with appropriate language model
+- Some engines (Google Vision, Azure) handle mixed languages automatically
+- Tesseract: specify multiple languages `tesseract input.png output -l eng+jpn+chi_sim`
+
+### Script-Specific Considerations
+| Script     | Direction | Special Handling                            |
+|------------|-----------|---------------------------------------------|
+| Latin      | LTR       | Standard processing                         |
+| Arabic     | RTL       | Right-to-left text flow, connected letters  |
+| CJK        | LTR/TTB   | May be vertical, no word spacing            |
+| Devanagari | LTR       | Connected headstroke (Shirorekha)           |
+| Thai       | LTR       | No spaces between words, requires segmentation |
+
+## Scanned vs Native PDF Processing
+
+### Quality Assessment
+```
+Native PDF (digitally created):
+- Text layer present: can extract directly without OCR
+- Perfect accuracy: text is already encoded
+- Structured: may contain tagged PDF elements
+- Detection: check for extractable text layer first
+
+Scanned PDF (image-based):
+- No text layer: requires OCR
+- Quality varies with scan settings
+- May have skew, noise, artifacts
+- Detection: if text extraction yields empty/garbage, it is scanned
+```
+
+### Preprocessing Pipeline for Scanned Documents
+
+#### 1. Deskew (Straighten)
+- Detect rotation angle via Hough Transform or projection profile
+- Rotate image to align text horizontally
+- Critical for accurate OCR - even 1-2 degree skew reduces accuracy
+
+#### 2. Denoise
+- Remove scanner artifacts, speckles, background texture
+- Apply Gaussian blur for mild noise
+- Use median filter for salt-and-pepper noise
+- Non-local means denoising for preserving text edges
+
+#### 3. Binarize (Convert to black/white)
+- Adaptive thresholding (Otsu's method) for uneven lighting
+- Sauvola binarization for documents with varying backgrounds
+- Critical step: poor binarization destroys thin strokes and small text
+
+#### 4. Additional Steps
+- **Border removal**: Crop black borders from scanning
+- **Page segmentation**: Identify text regions, images, tables separately
+- **Resolution check**: Ensure minimum 300 DPI for reliable OCR
+- **Contrast enhancement**: CLAHE (Contrast Limited Adaptive Histogram Equalization)
+
+### Decision Flowchart
+```
+1. Attempt text extraction from PDF
+2. If text extracted successfully -> use native text (no OCR needed)
+3. If no text or garbled text -> classify as scanned
+4. Assess scan quality (DPI, contrast, skew)
+5. Apply preprocessing pipeline
+6. Run OCR with appropriate engine
+7. Post-process: spell check, layout reconstruction
+```

package/data/free-skills/devflow-docs/template/agents/docs-specialist.md

@@ -0,0 +1,35 @@
+---
+name: docs-specialist
+description: Use this agent for document processing workflows — parsing client documents, generating deliverables, and managing document-based requirements.
+tools: Glob, Grep, LS, Read, Edit, Write
+---
+
+You are a Document Processing Specialist with expertise in extracting structured data from unstructured documents and generating professional deliverables.
+
+## Core Responsibilities
+
+### 1. Document Intake & Parsing
+- Activate appropriate skill based on document type
+- Extract structured data from PDF, Excel, PowerPoint
+- Handle multi-format document sets (RFP = PDF + Excel + Slides)
+
+### 2. Content Analysis
+- Cross-reference extracted data across documents
+- Identify gaps and inconsistencies
+- Flag ambiguous or missing information
+
+### 3. Document Generation
+- Activate `doc-generator` skill
+- Compile analysis outputs into professional documents
+- Ensure formatting meets client standards
+
+## Error Recovery
+- If document unreadable: report format/encoding issue, suggest alternatives
+- If extraction incomplete: flag missing sections, provide partial results
+- Max retries: 2 per document processing attempt
+
+## Checkpoint Protocol
+After processing each document:
+1. Save extracted content
+2. Log: document type, pages/sheets processed, extraction quality
+3. If interrupted: resume from last processed document

package/data/free-skills/devflow-docs/template/workflows/docs-workflow.md

@@ -0,0 +1,70 @@
+# Document Processing Workflow
+
+End-to-end document processing from intake through analysis to deliverable generation.
+
+## Phase 1: Intake & Classification
+
+1. **Document Receipt**
+   - Receive document(s) from user
+   - `docs-specialist` agent classifies each document by type
+   - Identify processing order based on dependencies
+
+2. **Type Routing**
+   | Document Type | Route To |
+   |---------------|----------|
+   | PDF (RFP, contract, spec) | `pdf-processor` skill |
+   | Excel (budget, resource plan) | `excel-analyzer` skill |
+   | PowerPoint (presentation, deck) | `presentation-reviewer` skill |
+   | Multi-format set | Process in order: PDF → Excel → PowerPoint |
+
+## Phase 2: Extraction (Parallel where possible)
+
+3. **PDF Processing** (if applicable)
+   - `docs-specialist` activates `pdf-processor` skill
+   - Extract text, tables, forms, and images
+   - Handle multi-column layouts and OCR if needed
+   - Output: Structured content with section mapping
+
+4. **Excel Analysis** (if applicable, parallel with step 3)
+   - `docs-specialist` activates `excel-analyzer` skill
+   - Parse budgets, resource plans, timelines
+   - Extract formulas and calculated fields
+   - Output: Structured data with analysis notes
+
+5. **Presentation Review** (if applicable, parallel with steps 3-4)
+   - `docs-specialist` activates `presentation-reviewer` skill
+   - Extract key decisions, action items, requirements per slide
+   - Identify diagrams and visual content
+   - Output: Slide-by-slide content summary
+
+## Phase 3: Cross-Reference & Analysis
+
+6. **Content Synthesis**
+   - `docs-specialist` cross-references extracted data across documents
+   - Identify inconsistencies (e.g., budget vs. scope mismatch)
+   - Flag gaps and ambiguous information
+   - Output: Cross-reference analysis report
+
+## Phase 4: Generation (if requested)
+
+7. **Document Generation**
+   - `docs-specialist` activates `doc-generator` skill
+   - Input: Extracted + analyzed data from phases 2-3
+   - Compile into requested deliverable format
+   - Apply formatting and professional structure
+   - Output: Generated document ready for review
+
+## Phase 5: Review & Delivery
+
+8. **Quality Check**
+   - Verify extraction completeness (all sections processed)
+   - Validate cross-references are accurate
+   - Confirm generated documents meet format requirements
+   - Present results to user with confidence indicators
+
+## Quality Gates
+
+- **After Phase 1**: Document classification confirmed
+- **After Phase 2**: Extraction quality ≥ 90% (based on section coverage)
+- **After Phase 3**: All cross-reference inconsistencies flagged
+- **After Phase 4**: Generated document matches requested format

package/data/free-skills/devflow-postproject/manifest.json

@@ -0,0 +1,13 @@
+{
+  "name": "@skillstore/devflow-postproject",
+  "version": "1.0.0",
+  "skills": [
+    "handover",
+    "maintenance-support",
+    "retrospective",
+    "client-feedback"
+  ],
+  "freeSkills": ["retrospective"],
+  "commands": ["postproject", "postproject/handover", "postproject/retro", "postproject/support"],
+  "modes": ["plugin", "template"]
+}

package/data/free-skills/devflow-postproject/plugin/commands/devflow.md

@@ -0,0 +1,34 @@
+---
+description: DevFlow smart router — automatically routes tasks to the right post-project skill
+argument-hint: [task description]
+---
+
+## Your Mission
+
+<task>
+$ARGUMENTS
+</task>
+
+## Smart Routing
+
+### Route Detection
+| Keywords/Patterns | Route To |
+|-------------------|----------|
+| handover, transition, KT, knowledge transfer, documentation | `handover` skill |
+| SLA, maintenance, incident, hotfix, support, on-call | `maintenance-support` skill |
+| retrospective, retro, lessons learned, improvement, feedback (internal) | `retrospective` skill |
+| client feedback, satisfaction, CSAT, NPS, upsell | `client-feedback` skill |
+| full transition, project closure, end-to-end | Full orchestration (see below) |
+
+### Full Orchestration
+1. `retrospective` (while context is fresh)
+2. `handover` (documentation + KT)
+3. `client-feedback` (satisfaction assessment)
+4. `maintenance-support` (ongoing support setup)
+
+### Ambiguity Handling
+If unclear: ask ONE clarifying question.
+
+## References
+- Load `rules/development-rules.md` for quality standards
+- Follow `protocols/error-recovery.md` for failure handling

package/data/free-skills/devflow-postproject/plugin/commands/postproject/handover.md

@@ -0,0 +1,21 @@
+---
+description: Create comprehensive handover documentation
+argument-hint: [project name or context]
+---
+
+## Your Mission
+
+Create handover documentation for:
+
+<project>
+$ARGUMENTS
+</project>
+
+## Process
+
+1. Activate the `handover` skill
+2. Inventory all project artifacts
+3. Create architecture and setup documentation
+4. Document known issues and pending items
+5. Plan knowledge transfer sessions
+6. Produce complete handover package

package/data/free-skills/devflow-postproject/plugin/commands/postproject/retro.md

@@ -0,0 +1,21 @@
+---
+description: Facilitate project retrospective
+argument-hint: [project name or sprint]
+---
+
+## Your Mission
+
+Facilitate retrospective for:
+
+<context>
+$ARGUMENTS
+</context>
+
+## Process
+
+1. Activate the `retrospective` skill
+2. Select appropriate retrospective technique
+3. Gather data on what went well and what didn't
+4. Generate insights and patterns
+5. Create action items with owners
+6. Produce lessons learned document

package/data/free-skills/devflow-postproject/plugin/commands/postproject/support.md

@@ -0,0 +1,21 @@
+---
+description: Setup maintenance and support framework
+argument-hint: [project or service context]
+---
+
+## Your Mission
+
+Setup maintenance support for:
+
+<context>
+$ARGUMENTS
+</context>
+
+## Process
+
+1. Activate the `maintenance-support` skill
+2. Define SLA tiers and response times
+3. Setup incident management process
+4. Create escalation matrix
+5. Define hotfix workflow
+6. Produce SLA document and operational runbook

package/data/free-skills/devflow-postproject/plugin/commands/postproject.md

@@ -0,0 +1,32 @@
+---
+description: Post-project workflow — handover, maintenance setup, retrospectives, and client feedback
+argument-hint: [task description]
+---
+
+## Your Mission
+
+<task>
+$ARGUMENTS
+</task>
+
+## Workflow
+
+Route to the appropriate post-project skill:
+
+1. **Understand the request** — What post-project activity is needed?
+2. **Route to skill**:
+   - Documentation/transition → Activate `handover` skill
+   - SLA/incidents/hotfix → Activate `maintenance-support` skill
+   - Lessons learned/improvement → Activate `retrospective` skill
+   - Client satisfaction/upsell → Activate `client-feedback` skill
+3. **Full post-project cycle** (if requested):
+   - Start with `retrospective` while project context is fresh
+   - Then `handover` for documentation
+   - `client-feedback` for satisfaction assessment
+   - `maintenance-support` for ongoing support setup
+
+## Sub-commands
+
+- `/postproject:handover [project]` — Create handover documentation
+- `/postproject:retro [project]` — Run project retrospective
+- `/postproject:support [context]` — Setup maintenance support