foundations
claudeWhat every project needs regardless of stack: decision records that people and agents can read, and releases with a changelog and a tag on main.
31 skills
Improve the harness your agents run on, one small verified change at a time: instruction files, skills, memory, hooks and permissions, repository checks, and your own prompts, ranked by the evidence of real sessions and by your goals.
Plugin harness-engineering from marketplace theultimate-dev in theultimate-dev/skills. Commands read from the repo's manifest — 5 days ago.
decision-recordsKeeps a project's decision log: short decision records (ADRs) in a decisions/ directory with an indexed README linked from AGENTS.md or CLAUDE.md. Use when the user settles a trade-off, asks to record or write up a decision, asks why something was chosen, mentions ADRs or architecture decisions, or
release-processMaintains a project's CHANGELOG.md in Keep a Changelog form and cuts releases: picks the SemVer bump from Conventional Commits, moves Unreleased into a dated version section, bumps declared version files, commits, tags vX.Y.Z on main, pushes, and relies on a GitHub Action to publish the GitHub Relea
marginal-gainsReviews a working session with a coding agent and proposes the two or three smallest changes to the user's harness that would pay off most: instruction files such as AGENTS.md or CLAUDE.md, skills, memory, hooks, permissions and settings, repository checks, and the user's own prompts. Reads the sess
curating-product-requirementsMaintains the product layer of a product context repository: overview, glossary, actors and use cases, and a requirement catalogue with permanent capability-prefixed IDs, priorities, target versions and statuses, plus non-functional requirements with stated targets. Integrates product briefs, produc
designing-system-architectureDesigns and evolves the architecture in a product context repository: C4 context and container views, components where they pay, runtime scenarios, data, deployment, cross-cutting concepts, and the versioned contracts several bets share, each section tagged with the version it arrives in and whether
planning-product-roadmapsPlans a product's versions and milestones in a product context repository: the gates between versions, version briefs with outcome, done-when, scope in and out, milestones as release slices with observable exit criteria, requirement coverage per milestone derived from the catalogue, prerequisites ou
reconciling-product-contextFolds delivered work back into a product context repository and keeps it coherent. Closes a bet after its pull requests merge: confirms every acceptance criterion was observed on merged code, updates requirement statuses and the architecture as built, writes contract details settled during implement
shaping-betsShapes a milestone of a product context repository into bets: epic-sized, appetite-bounded work units with an observable outcome, the requirement IDs they cover, acceptance criteria, no-gos, rabbit holes, cuttable scope, the contracts and decisions they rely on, and an ordered context pack. Checks t
structuring-product-contextCreates or adopts a product context repository: a Markdown repository, kept apart from the code, that holds a long-running product's whole context in layers sorted by how often they change (product, architecture, decisions, a roadmap with bets, design, research), with an AGENTS.md map, folder indexe
building-html-prototypesBuild responsive UI sketches and interactive prototypes as offline HTML: by default one self-contained file with embedded CSS, assets, and minimal JavaScript, or a design tree of pages that link the project's shared tokens and component styles by relative path. Use to compare design directions or re
curating-design-systemsKeep a project's design home, the long-lived folder of shared design foundations such as tokens.css and components.css with a folder per platform and per view, to its four layout rules, and check after every change that each page's links, sources and CSS url() references resolve inside it. Use when
designing-ux-flowsTranslate product requirements into UX flows, information hierarchy, screen responsibilities, interaction rules, content needs, and relevant interface states. Use when a user journey or feature behavior needs to be worked out before visual design or implementation.
exploring-visual-directionsDevelop and compare distinct visual direction briefs for a user-facing application or feature, including composition, typography, spacing, density, color, imagery, and motion, for a responsive page or a native platform surface such as a menu-bar panel, desktop window, phone widget, watch face, TV or
guiding-product-discoveryGuide a rough product, application, feature, or proof-of-concept idea through requirements, UX flows, visual directions, offline HTML prototypes, and a product/design handoff. Use to coordinate discovery across phases or resume unfinished design work; individual artifact requests belong to the relev
preparing-implementation-handoffsConsolidate requirements, UX flows, selected visual decisions, HTML prototypes, acceptance criteria, and suggested delivery slices into a product/design handoff. Use when discovery artifacts need to become a coherent brief for engineering planning, without inventing architecture or API contracts.
shaping-product-briefsTurn a rough application, feature, or proof-of-concept idea into a product brief with audience, outcomes, scope, prioritized functional requirements, assumptions, and acceptance criteria. Use before design or implementation when the problem and intended behavior need clarification.
brainstorming-solutionsResearches the codebase for prior art, reusable utilities, and constraints, generates two to four genuinely distinct technical approaches to an agreed spec.md, compares them in a table of architecture fit, risk, effort, and reversibility with a clear recommendation, and writes the approach the user
defining-verificationAgrees with the user how the agent will prove its own work before implementation starts (stage 3 of the product-engineering loop, gate G3 Verification). Discovers the project's test, lint, typecheck, build and CI commands, how to install, seed and start the app, and which browser-driving and other v
implementing-plansImplements one PR slice, a whole plan or one of its phases, from plans/NN-slug.md at stage 5 of the product-engineering loop: checks the assignment against spec.md, branches from a freshly synced base, writes a failing check first, makes the smallest coherent change, commits per logical step, runs t
planning-implementationPlans the implementation of an agreed spec.md at stage 4 of the product-engineering loop and runs gate G4, whose approval launches autonomous work. Decomposes the work item into plans and phases, chooses PR slices that each stay green and releasable, sets dependencies and parallel work, predicts eac
reviewing-code-changesReviews a code change through five lenses (architecture, security, conventions and idioms, efficiency, intent), each run by a fresh-context reviewer that forms its own view, then deduplicates and substantiates the findings, rates them blocking, should-fix or nit, and gives an APPROVE or REQUEST CHAN
running-implementation-loopsEntry point of the product-engineering loop: takes a work item from request to merged pull requests. Classifies it into a track (quick fix, bugfix, small change, feature, new app, refactor, spike), runs gates G1–G4 with the user, then runs each PR slice through implement, verify, review, PR and merg
shipping-pull-requestsOpens, categorizes and lands one pull request per PR slice at stages 8 and 10 of the product-engineering loop. Pushes the branch, opens the PR with the verification report, computes the human or agent review category from the actual diff, watches required CI on the current head, runs the full review
specifying-work-itemsInterviews the user to turn an issue, product brief, product-design handoff, or a bet from a product context repository into an agreed spec.md with numbered requirements (R1…), observable acceptance criteria (AC1…), non-goals, constraints, and assumptions. Grounds itself in the code first so it neve
verifying-implementationVerifies a committed candidate for real, without the user (stage 6 of the product-engineering loop). Checks out the handed-over branch, confirms HEAD is the handed-over commit SHA and never commits, starts the app itself from that checkout with docs/product-engineering/verification-profile.md, runs
building-prototype-betsBuilds an approved prototype bet in a loop until intent and result match: reads the bet at the commit it was approved at, plans, implements and commits each step as it finishes within the autonomy line, verifies every acceptance check by its named method with quoted evidence, has an independent revi
closing-prototype-betsCloses a prototype bet after the person has tried it: asks keep or remove at Gate 2, showing the build log's checks with evidence, the checks only the person's eye can judge, and the bet's keep-or-remove signal. Keep folds the result into the context: requirement rows become built, planned architect
framing-prototype-ideasFrames a prototype idea before any bet exists: checks what the prototype context and code already settle, compares two or three options, recommends one with a signal for removing it after a try, and gives a verdict of shape, park or stop in a one-page dated research note. Use when the user says I ha
shaping-prototype-betsShapes one prototype bet from a framed idea: an observable outcome, an appetite of half a day to two days, requirement deltas with reserved permanent IDs, acceptance checks that each name how they are verified, no-gos, rabbit holes, cuttable scope, a keep-or-remove signal, a context pack, and an aut
structuring-prototype-contextCreates or adopts a prototype context: a context/ folder beside a prototype's code that keeps product, architecture, decisions, bets and notes in a light profile where ideas are tried and then kept or removed, with an AGENTS.md map, derived indexes, a decision record for the graduation trigger and a
improving-promptsRewrites a rough prompt into an engineered one. Classifies where the prompt will run (chat request, system prompt, coding or research agent task, agent operating instructions, long-document job, extraction or classification), diagnoses its gaps, applies the prompt and context engineering techniques
What every project needs regardless of stack: decision records that people and agents can read, and releases with a changelog and a tag on main.
31 skills
A long-running product's whole context in a Markdown repository kept apart from the code: requirements with permanent IDs, a living architecture with shared contracts, decisions, and a roadmap cut into appetite-bounded bets that the product-engineering loop builds, folded back into the documents after every bet.
31 skills
From rough ideas to product/design handoffs: requirements, UX flows, visual directions, and offline HTML prototypes, as single files or as a design tree that shares the project's tokens and components, with an optional discovery coordinator.
31 skills
One implementation loop from a work item to merged pull requests: you are interviewed, pick an approach, agree the verification method and approve the roadmap; the agent then implements, verifies in the running app, reviews with five parallel lenses, and merges the PRs it may merge.
31 skills
A prototype's context beside its code and a fast loop for each idea: frame it, shape it into a bet you approve, build it with self-verification and an independent review, then try it and keep it or remove it, with history in the app's changelog.
31 skills
From a rough prompt to an engineered one: destination and gap diagnosis, a catalog of prompt and context engineering techniques with apply and reject criteria, dated model-family notes, and a copy-paste-ready result with the reasoning behind it.
31 skills
Matt Pocock's agent skills for real engineering (grilling, spec/ticket flows, TDD, code review, domain modelling and more).
39 skills
Persistent memory for Claude Code: the next session starts where the last one left off.
29 skills · 1 MCP server · hooks
Cross-session memory and token savings for coding agents.
6 skills · 1 agent · 1 MCP server · hooks
Up-to-date documentation lookup. Pull version-specific documentation and code examples directly from source repositories into your coding agent's context.
1 MCP server
Build powerful, stateful CLI interfaces for any GUI application using the cli-anything harness methodology.
5 commands
Automates browser interactions for web testing, form filling, screenshots, and data extraction
12 skills
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.