Menu
SkillsMP has collected 41 skills from Harduex/skills. Open a skill to review its source and details.
Showing 40 of 41 collected skills.
The universal base layer of standards for everything a project authors, code and prose alike, derived from the ASD-STE100 controlled-language specification: a controlled naming vocabulary with a per-project glossary, small imperative units with guard…
Converts an internal artifact — a design spec, a shipped branch or feature, an API or data contract — into a self-contained handoff document an external team can read fast and act on without any internal context. Use when handing a contract, integration…
Distributes uncommitted working-tree changes back to their originating commits on a multi-commit branch using git commit --fixup and a non-interactive autosquash rebase. Use when the user wants to "distribute" fixes into existing commits, clean up a branch…
Reviews code changes for correctness, security, test coverage, and design quality, producing severity-tagged findings. Supports local changes, specific commits, or GitLab merge requests. Use when user requests a code review, asks to review a diff, branch,…
Create new agent skills with proper structure, progressive disclosure, and bundled resources. Use when user wants to create, write, or build a new skill, or to update, audit, or modernize an existing skill's structure, frontmatter, or description.
Functional programming principles for structuring code as Data, Calculations, and Actions — immutability disciplines, stratified design (layers, abstraction barriers, minimal interfaces), duplication-removing refactorings and functional tools, timeline rules…
Drives a whole feature from a short prompt to merge-ready with minimal supervision — derives a machine-checkable done-contract, plans and executes tasks in a build→test→verify cycle, and persists all state to files on disk so the work survives context…
Produces a self-contained handoff prompt capturing the current session's state — goal, decisions, progress, key files, next steps — to paste into a fresh session and continue seamlessly. Use when the context is getting long, before compacting, or when the…
Use when a design, spec, or integration carries unknowns about an external or unfamiliar system — another team's codebase, service, API, wire format, schema, or closed package — and before asking the owning team anything. Also when TBDs block finalizing a…
Authors concise, reviewable design specs (written before implementation) and as-built architecture decision records (ADRs, written after a feature ships) for features and systems. Enforces an assertive, contract-first style throughout, plus a structure…
Condense a spec, design doc, RFC, ADR, or plan into a one-screen bullet overview that carries the actual decisions — what the system does, how, what it deliberately does not do, and what is still undecided. Use when asked to summarize or give an overview of a…
Create consistent, readable architecture, system, and flow diagrams using C4 conventions rendered in ASCII (for specs, RFCs, design docs, READMEs, PRs). Use when drawing or editing an architecture / system / context / container diagram, a change-footprint or…
Architects new features and systems with rigorous requirements interrogation before proposing solutions. Refuses to proceed with vague specs — demands precise constraints scaled to scope. Use when designing a new feature or system, evaluating architectural…
Record a short, caption-annotated screen video that proves a change works, instead of a written test report. Use when asked to film/record/screen-capture a verification run, to show a change working, or to produce a walkthrough a reviewer can watch; also use…
Orchestrates a multi-pass, multi-agent review→fix→verify loop over a branch, diff, or feature — parallel lens reviews produce severity-tagged findings, the most serious are fixed (one commit each), the work is re-verified, and the loop repeats until clean.…
Use before claiming any implementation is done, fixed, complete, or passing — and before committing or handing work over for testing. Green typecheck/lint is not verification; the changed behavior must be exercised and the evidence shown.
Validates and addresses reviewer feedback on the current branch's merge/pull request — comments fetched from GitLab or GitHub via their CLIs (glab, gh), or pasted into the chat. Use when asked to fetch, read, or check MR/PR review comments, validate whether…
Master router for the Evolving Harness framework. Consult this skill at the start of ANY non-trivial software task - a new feature, refactor, bug fix, migration, or raw idea - to route the work through the correct lifecycle skills with explicit approval…
Scaffolds or retrofits a project's agent harness — the AGENTS.md instruction file plus a standard set of per-domain project skills (coding standards, design system, data layer, db ops, e2e tests, browser-testing wrapper, domain playbooks). Use when setting up…
Sets up the answer-formatting style defined in output-style.md at the level the user picks — user-wide, current project, or this session only. Purely cosmetic — never alters reasoning, tool usage, file edits, or the factual content of the work.
Audits a git repo's full commit history for anything that makes it unsafe to publish — leaked secrets/credentials, PII, work identity, machine-local paths, and accidentally-committed binaries, DB dumps, or key material. Produces a safe-to-publish / not…
A structured deep-reasoning protocol distilled from the documented working methods of history's most exceptional thinkers (Einstein, Feynman, Darwin, Shannon, Poincaré, von Neumann) and from cognitive-science research on creative breakthrough. It replaces the…
Systematically diagnoses and resolves software defects through reliable reproduction, the scientific method, evidence-based instrumentation, and root-cause analysis. Use when investigating bugs, diagnosing errors, tracing unexpected behavior, or resolving…
Designs user experiences — user flows, wireframes, interaction states, component behavior specs, and UX critique. Use when mapping a user journey, wireframing a screen, defining a component's interaction/empty/error states, evaluating usability, or critiquing…
Use when the user has a vague or complex idea that needs validation before building. Turns rough ideas into tested, simplified specs and implementation plans through iterative assumption-testing. Triggers on phrases like "I have an idea", "I want to explore",…
Defines product requirements, breaks down work into actionable tasks, and prioritizes based on business value. Writes user stories with acceptance criteria. Use when writing an implementation plan, breaking a feature into ordered tasks with acceptance…
Analyzes any static website codebase, maps out the site structure, and guides the user through a page-by-page review to implement corrections, content updates, and improvements. Use when reviewing, refining, auditing, or updating an existing static site page…
Rigorous requirements interrogation methodology that refuses to proceed with vague specifications. Demands precise constraints before implementation. Applies conditional depth based on scope. Use when told "zero assumptions", "don't assume", or "you assumed…
Builds a self-contained interactive HTML node-link graph of any system — every component becomes a node clustered by layer, with typed, colour-coded edges for how the parts connect and talk to each other. Use when the user wants to visualise, map, or explore…
Use when reconstructing or documenting how something was built across past Claude Code agent sessions — mine the ~/.claude/projects transcripts (JSONL) for a feature, branch, or Jira ticket spanning many sessions and repos to produce an implementation…
Turn a workflow you just ran — with the corrections that shaped it — into a reusable skill. Use when you ask whether a good session is worth a new skill or an update, when extracting a repeatable process from past sessions, or when deciding if a learning…
Capture durable lessons from the current session — user preferences, project conventions, gotchas, and reusable techniques — and route each to its right home (the memory system, project/global CLAUDE.md or AGENTS.md, or an existing/new custom skill). Use when…
Use when asked to discover and analyze quick wins — high-value, low-effort improvements that ride the existing infrastructure — across a product or codebase: a "find quick wins / low-hanging fruit / what can we improve cheaply" scan spanning UX, architecture…
Use when rewriting a multi-commit branch into clean, single-purpose commits that each build on their own — linearize, curate, or clean up branch history, make every commit a self-contained working milestone, split an overloaded commit, reorder commits so…
Produces a report of shipped work for a given period or quarter (Q{N}) by mining merged git history across one or more repos, grouping commits by issue ticket into features, and attributing them to developers. Use when asked to summarize or audit what shipped…
Rigorously verifies the correctness, safety, or security of code, database migrations, authorization/permission rules, configuration, and business logic. Use this skill WHENEVER the user wants to check, audit, verify, or prove something about a system — for…
Generates a comprehensive, prioritized test plan for a feature, change, or system area — enumerating cases across happy paths, negative and boundary cases, state and lifecycle, permissions, cross-context parity, security, observability, and non-functional…
Conducts autonomous, multi-step deep research on complex topics using iterative planning, systematic exploration, source synthesis, and verified output generation. Use when the user asks for deep research, comprehensive analysis, literature review, technical…
Audits the AI coding agent ecosystem and compares against a project's current agent infrastructure. Researches Claude Code, Cursor, Codex, Gemini CLI, and other tools for new features, skill standards, and best practices, then produces a gap analysis with…
Audits and optimizes a collection of agent skills for compliance, concision, and cross-tool portability. Applies Agent Skills spec, Anthropic best practices, and progressive disclosure principles. Use when reorganizing skills, auditing skill quality, or…