# Matt Pocock study notes

> Independent, unofficial, non-commercial notes. Original works belong to their authors.

## Pages

- [Corpus](https://matt.withkeya.com/en/corpus.md): Search summaries written for this site and follow links to the public originals.
- [Learning from Matt](https://matt.withkeya.com/en/learn-from-matt.md): Choose a starting point around a problem you actually need to solve.
- [Core Ideas](https://matt.withkeya.com/en/core-principles.md): Turn an agent failure into an environment problem you can improve and verify.
- [Three Lines for Practice](https://matt.withkeya.com/en/punch-lines.md): Give a memorable line a scope, a limit, and a practical next step.
- [Workflow Overview](https://matt.withkeya.com/en/ai-coding-workflow.md): Choose enough process for the risk and uncertainty of the change.
- [Alignment and Prototypes](https://matt.withkeya.com/en/alignment-and-prototypes.md): Separate facts, decisions, and experience questions to agree on observable behavior.
- [Specs and Tickets](https://matt.withkeya.com/en/specs-and-tickets.md): Capture settled decisions and slice them into independently verifiable tickets.
- [Implementation and Feedback](https://matt.withkeya.com/en/implementation-and-feedback.md): Use repeatable failure and passing evidence to decide whether a change is complete.
- [Review & PR](https://matt.withkeya.com/en/review-and-pr.md): Give reviewers separate findings, observable behaviour and a clear account of merge risk.
- [Context & handoff](https://matt.withkeya.com/en/context-and-handoff.md): Preserve the state a fresh session needs to verify progress and continue.
- [Agent documents](https://matt.withkeya.com/en/agent-documents.md): Turn a crowded instruction file into useful pointers with explicit triggers.
- [Standards & checks](https://matt.withkeya.com/en/standards-and-checks.md): Automate rules with definite answers and reserve judgement for review.
- [Decisions & retrospectives](https://matt.withkeya.com/en/decisions-and-retrospectives.md): Keep durable reasoning and use session evidence to propose a small environmental improvement.
- [TypeScript foundations](https://matt.withkeya.com/en/typescript-foundations.md): Validate unfamiliar data, then preserve what you have learned through types.
- [Practical type tools](https://matt.withkeya.com/en/practical-type-tools.md): Choose a type tool for a concrete mistake and explain what it leaves unresolved.
- [States and Module Boundaries](https://matt.withkeya.com/en/state-and-module-boundaries.md): Make invalid states harder to express and module behaviour easier to depend on.
- [Writing to Teach](https://matt.withkeya.com/en/writing-to-teach.md): Begin with a reader’s concrete problem, then check learning with a new problem.
- [Feedback and Attribution](https://matt.withkeya.com/en/feedback-and-attribution.md): Turn reproducible failures into improvement candidates and keep contributions traceable.
- [Open-source Guide](https://matt.withkeya.com/en/open-source-guide.md): Choose a repository for a concrete problem, then check its scope and licence.
- [Cheat Sheets](https://matt.withkeya.com/en/cheat-sheets.md): Find the next action by its trigger, with a route back to the full explanation.
- [Learning Paths](https://matt.withkeya.com/en/learning-path.md): Use three diagnostic questions to choose a route, then assess progress through checkable work.
- [About This Site and Copyright](https://matt.withkeya.com/en/about-and-copyright.md): An independent, unofficial collection that distinguishes quotations, interpretation and licensed adaptations.