Requirement Compass
Extract goals, scope, constraints, and acceptance
AI agents, from intent to delivery.
I connect Skills, MCP, and specialist tools across requirements, engineering, documentation, and verification to create usable results.
From intent to delivery
Clarify
Define outcomes, constraints, and acceptance
How I use AI at work
Engineering
From codebase discovery and architecture to implementation, testing, and deployment, keeping requirements, implementation, test results, and final behavior aligned.
- Codebase impact analysis
- Java / Python / TypeScript
- APIs, databases, automation
- Tests, builds, runtime checks
- Docker and fault diagnosis
Project delivery
Using AI for scope, tasks, schedule, risks, quality, and acceptance so project materials are more complete and traceable.
- Requirements and WBS
- Milestones and dependencies
- Minutes and action items
- Acceptance and quality gates
- Reports and retrospectives
Knowledge engineering
Turning project material and technical experience into searchable, reusable assets for RAG, engineering collaboration, and continuous delivery.
- Project material structuring
- Skills / SOP / templates
- RAG corpora and knowledge tags
- API, architecture, and operations docs
- Decision records and retrospective knowledge
Explore capabilities and tools
Skills catalog20
Scope Decomposer
Break complexity into tasks, dependencies, priorities, and milestones
Architecture Scout
Understand structure, debt, and impact
Code Builder
Implement Java, Python, and TypeScript outcomes
Refactor Guard
Improve local structure while preserving behavior
API Contract
Design interfaces, schemas, errors, and integration notes
Data Pipeline
Clean, transform, batch, log, and isolate exceptions
RAG Engineer
Build parsing, retrieval, citations, evaluation, and governance
Prompt Lab
Design instructions, context, formats, and robustness tests
Agent Orchestrator
Coordinate roles, parallel tasks, context, and integration
MCP Connector
Connect structured context and callable tools
Browser Operator
Navigate, read, act, and validate visible state
Desktop Workflow
Connect local apps, scripts, screenshots, and UI automation
Document Studio
Create requirements, designs, technical documentation, SOPs, and project reports
Spreadsheet Analyst
Organize data, formulas, charts, and checks
Visual QA
Use screenshots, rendering, and diffs for UI acceptance
Test Engineer
Design relevant tests and explain failures
Risk Sentinel
Identify security, access, quality, dependency, and delivery risks
Delivery Reporter
Turn progress, issues, decisions, and outcomes into management views
Retrospective Memory
Capture learning as Skills, templates, scripts, and next inputs
Tools and MCP
Codebase Explorer
Codebase and impact analysis
Browser Control
Web interaction and visible-state verification
Desktop Control
Local application automation
Document Studio
Word and structured documents
Spreadsheet Engine
Sheets, formulas, charts, checks
PDF Inspector
PDF generation and layout QA
Presentation Builder
Decks and management reporting
Image Studio
Original visuals and image work
Research Navigator
Research and option comparison
Project Workspace
Task, file, and delivery context
API Integrator
Model and business-service integration
Visualization Lab
Charts and interactive explanations
Runtime Bridge MCP
Connects local execution, code, and engineering tools so an Agent can move from analysis to action.
Knowledge Mesh MCP
Connects developer knowledge and structured professional context for stronger solutions and decisions.
Collaboration and verification
I avoid self-ratings and show real workflows, delivery scenarios, capability catalogs, and verification methods. Each item maps to a concrete action and usable output.
Agent orchestration
5-stage flow / Goals, context, roles, dependencies, and concurrency
Skills design & reuse
20-item catalog / Domain rules, templates, scripts, and acceptance
MCP & tool integration
2 channels / Runtime tools and structured knowledge context
Browser and desktop automation
Live UI checks / Page operations, visible state, scripts, and visual acceptance
Code, docs, and data delivery
3 delivery lanes / Engineering, documentation, tables, and visualization
Verification & review
4 check types / Builds, tests, rendering, data validation, and issue closure
These figures describe the current public workflow and catalog, not a rating, ranking, or external assessment.
Application scenarios
Complex engineering
Locate impact, implement incrementally, run relevant checks, and produce maintainable code and handoff.
AI project management
Turn goals, scope, dependencies, risks, and acceptance into executable delivery structures.
High-quality documents
Create and verify requirements, technical designs, project reports, SOPs, and knowledge documentation.
Data & automation
Handle sheets, logs, batch work, and repetitive flows with clear status, exception handling, and result checks.
Research & decisions
Synthesize authoritative research, compare options, and produce clear recommendations.
Quality & visual QA
Confirm delivery quality through builds, tests, screenshots, rendering, and diff checks.
Quality and delivery principles
Outcome alignment
Clarify the problem and desired result so every Agent action supports delivery.
Information quality
Use clear context and reliable material to keep analysis, solutions, and communication consistent.
Scope focus
Control task scope and change radius so effort stays on what matters most.
Quality assurance
Use tests, visual checks, and data validation to confirm the deliverable works.
Stable delivery
Keep practical rollback paths for important changes and releases.
Continuous improvement
Capture effective methods as Skills, scripts, templates, and project assets.
AI agents, from intent to delivery.
I bring Skills, MCP, browsers, and specialist tools into real work, coordinating requirements, engineering, documentation, and verification to produce results that can be inspected and used.
Agent orchestrationSkills workflowsMCP integrationMulti-agentBrowser automationCode & docsQuality verification
AI PRACTICE / CONTINUOUS COLLABORATION
Proving AI collaboration through inspectable workflows.
I avoid self-ratings and show real workflows, delivery scenarios, capability catalogs, and verification methods. Each item maps to a concrete action and usable output.
Delivery loop
From goal clarification and execution to verification and learning
Delivery focus
Engineering, project delivery, and knowledge engineering
Applied scenarios
Code, projects, documents, data, research, and QA
Verification
Builds, tests, rendering, and data checks
Skills catalog
Callable capability units listed on this page
MCP channels
Runtime and knowledge context connections
These figures describe the current public workflow and catalog, not a rating, ranking, or external assessment.
WORKFLOW COVERAGEAgent collaboration map
Scope
Code
Docs
Data
QA
Review
VERIFIABLE PRACTICEAgent work that can be inspectedEach capability maps to a workflow, catalog entry, or verification action shown on this page.
Agent orchestration5-stage flow
Goals, context, roles, dependencies, and concurrencySkills design & reuse20-item catalog
Domain rules, templates, scripts, and acceptanceMCP & tool integration2 channels
Runtime tools and structured knowledge contextBrowser and desktop automationLive UI checks
Page operations, visible state, scripts, and visual acceptanceCode, docs, and data delivery3 delivery lanes
Engineering, documentation, tables, and visualizationVerification & review4 check types
Builds, tests, rendering, data validation, and issue closureFLOW / OPERATING MODEL
Five-stage AI agent delivery flow
Clarify
Define outcomes, constraints, and acceptance
PARALLEL DELIVERY LANES
AI supporting engineering, project delivery, and knowledge systems in parallel.
Engineering
From codebase discovery and architecture to implementation, testing, and deployment, keeping requirements, implementation, test results, and final behavior aligned.
- Codebase impact analysis
- Java / Python / TypeScript
- APIs, databases, automation
- Tests, builds, runtime checks
- Docker and fault diagnosis
Project delivery
Using AI for scope, tasks, schedule, risks, quality, and acceptance so project materials are more complete and traceable.
- Requirements and WBS
- Milestones and dependencies
- Minutes and action items
- Acceptance and quality gates
- Reports and retrospectives
Knowledge engineering
Turning project material and technical experience into searchable, reusable assets for RAG, engineering collaboration, and continuous delivery.
- Project material structuring
- Skills / SOP / templates
- RAG corpora and knowledge tags
- API, architecture, and operations docs
- Decision records and retrospective knowledge
TOOLS / SKILLS / MCP
A three-layer Agent stack: capabilities, tools, and context.
Twenty specialized capability units form the Skills layer, 12 cross-modal extensions form the tool layer, and two core connections form the MCP context layer. I combine them around each objective, connecting code, documents, data, browsers, and visual verification into one delivery chain.
20Capability units12Cross-modal tools2Custom context channels
Skills layer
Twenty callable capability units built from stable practice
Rules, templates, scripts, references, and acceptanceMCP context layer
Two core channels connecting runtime and knowledge
Structured context without repetitive copy and pastePlugin tool layer
Twelve cross-modal extensions for delivery
Code, documents, sheets, browsers, visuals, and contentMulti-Agent
Run independent work in parallel and integrate centrally
Useful for research, implementation, review, and comparisonBrowser & Desktop
Operate and validate real pages and local applications
DOM, screenshots, UI automation, local scriptsLocal + API
Combine local models, DeepSeek, and engineering services
Balancing cost, speed, data security, and controlPERSONAL AGENT PLATFORM / 20 + 12 + 2
Twenty Skills form my Agent capability matrix.
Twenty Skills handle professional actions, 12 tool extensions provide cross-modal execution, and two MCP channels connect runtime and knowledge context. Together they support engineering, project delivery, content production, and quality assurance.
Explore the complete Skills and tools catalog
Requirement Compass
Extract goals, scope, constraints, and acceptance
Scope Decomposer
Break complexity into tasks, dependencies, priorities, and milestones
Architecture Scout
Understand structure, debt, and impact
Code Builder
Implement Java, Python, and TypeScript outcomes
Refactor Guard
Improve local structure while preserving behavior
API Contract
Design interfaces, schemas, errors, and integration notes
Data Pipeline
Clean, transform, batch, log, and isolate exceptions
RAG Engineer
Build parsing, retrieval, citations, evaluation, and governance
Prompt Lab
Design instructions, context, formats, and robustness tests
Agent Orchestrator
Coordinate roles, parallel tasks, context, and integration
MCP Connector
Connect structured context and callable tools
Browser Operator
Navigate, read, act, and validate visible state
Desktop Workflow
Connect local apps, scripts, screenshots, and UI automation
Document Studio
Create requirements, designs, technical documentation, SOPs, and project reports
Spreadsheet Analyst
Organize data, formulas, charts, and checks
Visual QA
Use screenshots, rendering, and diffs for UI acceptance
Test Engineer
Design relevant tests and explain failures
Risk Sentinel
Identify security, access, quality, dependency, and delivery risks
Delivery Reporter
Turn progress, issues, decisions, and outcomes into management views
Retrospective Memory
Capture learning as Skills, templates, scripts, and next inputs
12 PLUGIN EXTENSIONS
2 MCP CONTEXT CHANNELS
Connects local execution, code, and engineering tools so an Agent can move from analysis to action.
Connects developer knowledge and structured professional context for stronger solutions and decisions.
APPLICATION SCENARIOS
Representative applications
Complex engineering
Locate impact, implement incrementally, run relevant checks, and produce maintainable code and handoff.
AI project management
Turn goals, scope, dependencies, risks, and acceptance into executable delivery structures.
High-quality documents
Create and verify requirements, technical designs, project reports, SOPs, and knowledge documentation.
Data & automation
Handle sheets, logs, batch work, and repetitive flows with clear status, exception handling, and result checks.
Research & decisions
Synthesize authoritative research, compare options, and produce clear recommendations.
Quality & visual QA
Confirm delivery quality through builds, tests, screenshots, rendering, and diff checks.
QUALITY / GOVERNANCE
Improving speed, quality, and control together.
Outcome alignment
Clarify the problem and desired result so every Agent action supports delivery.
Information quality
Use clear context and reliable material to keep analysis, solutions, and communication consistent.
Scope focus
Control task scope and change radius so effort stays on what matters most.
Quality assurance
Use tests, visual checks, and data validation to confirm the deliverable works.
Stable delivery
Keep practical rollback paths for important changes and releases.
Continuous improvement
Capture effective methods as Skills, scripts, templates, and project assets.
NEXT / ROLE FIT