mattpocock/skills
Skills
This project is an AI agent workflow framework and development toolkit designed for AI-driven software engineering. It provides a system of modular instructions, prompt libraries, and standardized routines to orchestrate complex engineering sequences and automate the decomposition of plans into technical tasks.
The system differentiates itself through advanced context management and prompt engineering, using state compression and handoff documents to preserve conversation history between different AI sessions. It employs a structured library of prompt skills and high-signal trigger words to ensure predictable agent behaviors across modular capabilities.
The toolkit covers a broad surface of professional engineering practices, including test-driven development, domain modeling, and software architecture analysis. It incorporates requirement engineering through design stress testing and interview-based validation, as well as automated issue triage and workflow routing to manage the software development lifecycle.
Additional utilities provide version control guardrails to block destructive commands and automation for authoring interactive educational content.
Features
- Agent Context Management - Implements state compression and handoff documents to bridge conversation context between different AI sessions.
- AI Agent Orchestration Frameworks - Provides a framework for routing tasks, managing context handoffs, and orchestrating complex engineering sequences for LLMs.
- Agent Handoffs - Provides mechanisms to summarize conversation state into documents for seamless context transfer between specialized agents.
- Modular Capability Compositions - Defines discrete agent behaviors as structured prompt files that can be assembled into modular capabilities.
- Agentic Context Management - Implements a methodology for preserving session state and compressing conversations into handoff documents for agents.
- Agent Prompt Templates - Creates modular and standardized system prompt structures to ensure predictable agent behaviors.
- Agentic Workflow Automation - Designs structured sequences and routing logic to orchestrate AI agents through complex engineering tasks.
- Agent Technical Specifications - Produces structured technical specifications for issues to enable automated implementation by AI agents.
- AI Agent Skills - Defines modular, prompt-based capabilities that can be invoked by either human users or other AI models.
- AI Workflow Orchestrators - Directs users to the most appropriate skill or automation flow based on their goals and current situation.
- Conversation State Management - Compact active chat history into markdown handoff documents to bridge state between agent sessions.
- Multi-session Context Synthesizers - Bridges state between different chat sessions or prototypes by compacting conversations into markdown files.
- Software Engineering Prompt Libraries - Provides a library of prompts to guide AI agents through professional engineering practices like TDD and architectural refactoring.
- Prompt Libraries - Provides a structured library of prompt skills and modular capabilities to ensure predictable AI agent outputs.
- Task Decomposition Systems - Deconstructs high-level requirement documents into independent, vertical-slice issues that deliver end-to-end functionality.
- Work Management Systems - Tracks bugs, tasks, and requirements through a central system to coordinate development.
- AI Development Toolkits - Offers a system of modular instructions to automate issue triage, requirement synthesis, and task decomposition.
- Development Workflows - The Software Engineering system implements features or fixes bugs using a disciplined red-green-refactor loop.
- Vertical-Slice Decomposition - Deconstructs specifications and requirement documents into independent, vertical-slice issues.
- Issue Tracking Systems - Organizes bug reports and feature requests into structured tasks managed through a defined state machine.
- Triage State Machines - Moves tasks through a defined state machine of triage roles to manage project progression.
- Design Stress Testing - Iteratively challenges architectural plans through a structured questioning process to identify edge cases and gaps.
- AI-Assisted Development - Integrates AI agents into the software lifecycle using TDD and deep-module architecture patterns.
- Task Decompositions - The Software Engineering system transforms a project plan into independent issues that deliver end-to-end functionality.
- Requirement to Task Decomposition - Deconstructs high-level requirement documents into granular issues for agent implementation.
- Software Architecture Design - Applies deep-module principles to create small interfaces that manage complex internal behaviors.
- Deep Module Design - Structures code by hiding complex behavior behind minimal interfaces to improve maintainability.
- Design Stress Testing - Uses iterative interviewing and stress testing to refine architectural plans and resolve domain conflicts.
- Implementation Planning - Defines public interfaces and prioritizes critical behaviors based on project domain language.
- Test-Driven Development Loops - Implements a disciplined red-green-refactor cycle of writing failing tests before producing the minimal required code.
- Invocation Access Controls - Restricts whether specific agent skills are triggered manually by a user or automatically by a model.
- Skill Routing Indexes - Provides a central entry point that indexes and directs users to other manual agent skills.
- Prompt Engineering - Applies standardized vocabulary and structured formats to prompts to ensure predictable and reliable agent outputs.
- Trigger Word Optimization - Links prompts to specific agent behaviors using high-signal words to improve response reliability.
- Content Authoring Tools - Provides tools for authoring educational lessons by reusing existing components and extracting assets.
- Boundary Stress Testing - Probes boundaries between business concepts by inventing scenarios that force precision in the domain model.
- Domain Refinement - Refines project terminology and domain models by testing edge cases and updating documentation.
- Logic Validation - Compares stated business rules and conceptual relationships against the actual codebase to validate logic.
- Pull Request Workflows - Standardizes the movement of tickets through a state machine to categorize bugs and enhancements.
- Engineering Workflow Initializers - Automates the configuration of issue trackers, triage labels, and documentation layouts for engineering skills.
- Logic Prototypes - The Software Engineering system creates minimal interactive terminal applications to test state machines and business logic.
- Task Decompositions - The Workflow Automation system divides large skills by invocation trigger or sequence to prevent rush and improve task granularity.
- Educational Lessons - Produces self-contained interactive HTML lessons designed as pedagogical units for learners.
- Architectural Analysis - Scans for architectural friction and proposes refactors to improve module depth and structure.
- Architecture Decision Records - Records high-impact technical trade-offs as formal architectural decision records to provide long-term context.
- Post-Implementation Workflows - The Software Engineering system improves code structure and removes duplication after tests pass to ensure functionality.
- Design Decision Stress Testing - Implements a rigorous process for questioning design plans until every branch of the decision tree is resolved.
- Rapid Prototyping - The Software Engineering system creates throwaway terminal apps or UI variations to validate logic and design.
- Glossaries - Resolves naming conflicts by establishing a canonical project vocabulary to ensure consistency across documentation.
- Interview-Based Design Validation - Questions users about plans and designs to resolve decision branches and validate architectural assumptions.
- Project Requirement Specifications - Executes an iterative interview process to sharpen project ideas and validate technical designs.
- Manual Fix Verification Workflows - Implements workflows to reproduce reported bugs from user steps to validate that fixes perform their intended function.
- Product Requirements Generation - Synthesizes conversation history and codebase context into formal product requirement documents for issue trackers.
- Role-Based Triage State Machines - Moves tasks and issues through a predefined sequence of status roles to manage project progression.
- Content Scaffolding - Generates structured directory hierarchies and stub files for educational sections and problems from a plan.
- Iterative Design Reviews - Walks through architectural candidates to resolve constraints while iteratively updating domain models.
- Architecture Refinement Automation - Scans code for improvement opportunities and generates visual reports to guide architectural refactoring.
- Software Requirements Analysis - Synthesizes conversation history into formal product requirement documents.
- Structured Bug Diagnosis - Executes a structured loop of reproduction and hypothesis testing to resolve complex regressions.
- System Seam Definition - Identifies and places boundaries where behavior can be altered without modifying calling code.
- Testability Optimizations - Optimizes interfaces to accept dependencies and return results, reducing side effects to improve testability.
- Ubiquitous Languages - Establishes a consistent project glossary by resolving naming conflicts and recording canonical definitions.
- Workflow Routing - Routes users to specific engineering skills or workflows based on the development context and technical requirements.
- Workflow Scaffolding - Scaffolds per-repository settings for issue tracking and domain documentation.
- Workflow Task Definitions - The Workflow Automation system establishes ordered actions with checkable completion criteria to ensure thorough task completion.
- Stress Testing - Employs iterative user interviews to challenge assumptions and verify architectural plans before implementation.
- Agentic Development Frameworks - Curated library of reusable skills for enhancing agent capabilities.
- AI Skills - Skills for engineering-grade AI coding.
