APA
h0ffmann. (2026). skills [Computer software]. https://github.com/h0ffmann/skills
BibTeX
@software{h0ffmann_skills_2026,
author = {h0ffmann},
title = {skills},
year = {2026},
url = {https://github.com/h0ffmann/skills},
license = {MPL-2.0},
note = {ORCID 0009-0009-1056-7661}
}
Source: CITATION.cff
52 agent skills. A skill is one folder with a SKILL.md file.
A ⏸ skill runs only when you call it by name (tag manual).
awesome-skills lists other skill collections.
agents is the separate index of the Claude agents.
Categories: code (2) · dev-flow (11) · figures (4) · nix (4) · notes (3) · ouroboric (4) · research (5) · web (4) · writing (9) · ww3 (6) · elsewhere (12)
In a Claude Code session:
/plugin install waza --marketplace h0ffmann/skills
In a terminal:
claude plugin install waza --marketplace h0ffmann/skills
| Skill | What it does | Tags |
|---|---|---|
| lean-code | Write, change or review code with the smallest correct diff. Question the need, reuse what exists, prefer the standard library, add no speculative abstraction. Use on any coding task, or for "simplest" or "yagni". | codereview |
| loco | Write, review or debug a Loco app (loco-rs, Rails for Rust): controllers, Sea-ORM models, migrations, workers, tasks and mailers. Use for AppContext, cargo loco, or a Loco app on a NixOS host. | coderustweb |
| Skill | What it does | Tags |
|---|---|---|
| agentskills-checker | Check skill folders against the Agent Skills specification with a script: frontmatter, names, limits and file references. Use before a commit that adds or changes a SKILL.md, or to audit a skill collection. For skill quality, use skill-validator. | dev-flowskillslint |
| bipr ⏸ | Abre ou atualiza o PR do branch atual com título e corpo bilíngues, pt-BR primeiro e o espelho em inglês, gerados a partir dos commits. Use quando pedirem /bipr ou um PR bilíngue. | dev-flowgithubprpt-brmanual |
| easy-issues ⏸ | Find small open issues that are ready for an agent, and solve each one as its own tested PR. Works for members and for contributors through a fork. Use for /easy-issues. | dev-flowgithubissueprautomationmanual |
| handoff | Save and resume work sessions as handoff notes in an Obsidian vault, and keep one digest note for each project there. Use for "save this session", "write a handoff", "load my last context" or "what was I working on". Not for decisions (use mip) or conventions (use AGENTS.md). | dev-flownotesobsidianagents |
| mip | Write or revise an Improvement Proposal (MIP), a numbered design document for a non-trivial feature, integration or architecture change, written before the code. Use when the user says "MIP", "improvement proposal", "propose a feature" or "write up a design for". | dev-flowplanningdesign-doc |
| mip-overnight ⏸ | Unattended runner for MIP task lists. Works one ready task at a time, opens a Draft PR for each, and stops cleanly at a usage limit or a real blocker. Only when typed by the owner. | dev-flowplanningprautomationmanual |
| mip-tasks ⏸ | Turn an accepted MIP into an ordered task list and deliver it as small stacked PRs, one task for each branch and PR. Use for `/mip-tasks NNNN` or a restack. | dev-flowplanningprgithubmanual |
| pr-images ⏸ | Open or update a bilingual PR like bipr, with before and after images in its body. Use for each PR that changes an image, a drawing, a page or a screen, or for /bipri. | dev-flowgithubprpt-brmanual |
| skill-validator | Write, review, debug or evolve an agent skill. Use to create or edit a SKILL.md, improve a description, fix a missed or wrong activation, or test a skill before it ships. For the format check only, use agentskills-checker. | dev-flowskillsreviewtesting |
| triage ⏸ | Turn a raw idea, a voice note or a bug report into one issue body that is ready to claim. A human files it. Use for "triage this" or "make an issue out of this". | dev-flowgithubissueplanningmanual |
| voice-note | Transcribe a local voice note (a WhatsApp .ogg or similar) with local Whisper, write the transcript next to the audio, then anonymize it by hand. Use when a MIP, an issue or a task needs the words of a voice note. | dev-flowaudionotes |
| Skill | What it does | Tags |
|---|---|---|
| archify | Create polished, validated architecture, workflow, sequence, data-flow, and lifecycle/state diagrams as explorable standalone HTML with inline SVG, dark/light themes, optional trace motion, and PNG/JPEG/WebP/SVG/WebM export. Accept plain-language requirements or pasted Mermaid flowchart, sequenceDiagram, and stateDiagram input; inspect repository evidence when the diagram must reflect real code. Use when the user asks to visualize system architecture, infrastructure, cloud/security/network topology, technical workflows, API call sequences, request lifecycles, data pipelines, ETL/ELT, data lineage, state machines, or to convert/beautify Mermaid. | figuresdiagram |
| ddraw | Draw a one-line drawing or a brushed kanji as Desmos equations and as an SVG from the same equations. Examine the render and draw again until it is good. Use for a README drawing, a logo or art in Desmos. | figuresdiagramsvgdrawing |
| figure | Add, change or explain a diagram in this repo's Markdown (README.md, README.en.md, docs): a Mermaid fence or an Archify picture. Use when asked for a diagram, flowchart, mind map or 'a picture of' something, or to update one after its prose changed. | figuresdiagrammarkdownpt-br |
| svg-diagram | Draw an architecture or repository map as one hand-written SVG for a README: a dark grid card, one colour for each kind of part, a legend. Use when Mermaid is too plain. For a figure with a reading card, use figure. | figuresdiagramsvgmarkdown |
| Skill | What it does | Tags |
|---|---|---|
| nix-debug | Find why a Nix evaluation, build, activation, service or boot fails. Update inputs, pins, hashes and carried patches safely. Use for a failed nix flake check, build or deploy, a flake.lock bump, or a rescue from a live USB. | nixnixosdebug |
| nix-lab | Write or change a self-contained Nix flake lab: a toolchain or dev shell that other repositories use. Covers locks, Rust tools, a just 1.21 justfile and the local gate. Use for flake.nix, devShells or a new lab. | nixtoolchain |
| nixos-vps | Change the NixOS configuration of a small live VPS that a merge to main deploys (deploy-rs with magic rollback). Use for modules, services, nginx, ACME, users, the resource budget, flake updates and the deploy itself. Never deploys. | nixnixosdeploy |
| systemd-hardening | Harden a systemd service written in Nix (systemd.services.<name>) with sandboxing, capability limits and syscall filters, then prove it still works. Use when you write or review a NixOS service module. | nixsystemdsecurity |
| Skill | What it does | Tags |
|---|---|---|
| pirsig | Drive Pirsig (psg), the PARA and Zettelkasten second brain, and its Hippo rhizomes. Capture and link notes, file projects and tasks, ingest arXiv or RSS, sync Obsidian. Use for Pirsig, psg, Hippo or a fixation. | noteszettelkasten |
| pirsig-slips | Work with ideas on paper slips, as Pirsig did in Lila: one idea on each slip, sorted by hand. Use for slips to copy, photos of slips, "which comes first" or piles to name. | noteszettelkasten |
| zk-method | Check that notes follow the Zettelkasten method: kind, atomicity, own words, links with reasons, sources. Lint a Markdown vault and propose fixes. Use for "is this a good zettel" or "audit my notes". | noteszettelkastenlint |
| Skill | What it does | Tags |
|---|---|---|
| ponytail-review | > Quality review of a change: is the logic right, is it safe, does it hold under real load, is risky code tested, is it fast enough, and is every line needed. Reads the connected code, not only the diff. Each finding is explained in plain English. Use for "review this", "code review", "review the last commit", "review my PR", "is this over-engineered", /ponytail-review. | ouroboriccodereview |
| retro ⏸ | Do a retrospective of a coding session and propose changes to the agent's environment: checks, navigation, steering files and tool cost. Use for /retro. | ouroboricagentsreviewmanual |
| sharingan | Read, evaluate and adopt (REA) tools from another repository, with the licence of each item decided before a copy. Use for "port this", "copy this skill" or "sharingan <url>", and before you copy any file from another repository. | ouroboricagentscodelicence |
| writing-styles | Select the writing style of a document from its reader: agent-facing, procedure, spec, guide, reference, research, journal, public help or UI copy. Then apply the skill and rules of that style. Use before you write a new document. | ouroboricwritingenglish |
| Skill | What it does | Tags |
|---|---|---|
| citation-score | Score how much of a text is cited, from 0% to 100%, with a coloured verdict. Use to score the citations of a phrase, a text, a file or a URL. | researchcitation |
| cite-release | Prepare a citable release and keep its Zenodo and citation metadata correct. Use to release or tag a version, mint a DOI, add an author, or change .zenodo.json, CITATION.cff or the citation block. | researchcitationrelease |
| eli5 | Explain a topic of the current project to a reader who knows nothing about it. Draw a picture first and take each fact from the checkout. Use for /eli5 <topic> or "explain this simply". | researchexplaindiagram |
| open-text | Get the full text of a work that is free to copy, keep it byte for byte with its licence, and index it. Use to fetch a book, story, essay or poem. When the text is not free, it gives a citation only. | researchlicence |
| paper | Write, check and prepare a citable research text (paper, proposal or report). Each claim has a mark, and a log says how each source was checked. Use for a draft, a new section or reference, or a check before a PR. | researchcitationwriting |
| Skill | What it does | Tags |
|---|---|---|
| citizen-science | Keep a public open-data site checkable and easy to join - sources, freshness, method, limits, privacy, licence and how to contribute visible on each page. Use when you add or review a page, a section or a panel that shows data to the public. | webopen-data |
| maps | Show areas, points and data fields on a web map with MapLibre and free tiles first, or Mapbox GL when a project already uses it. Covers CAP alert areas to GeoJSON, data-driven styling, performance and token safety. Use for any map on a page, an app or an image. | webmaps |
| news-post | Write, change or review a news post in pt-BR and English. Each claim has a source and the build calculates the reading time. Use for a post on a news page, or when somebody asks about the reading time of a post. | webwritingpt-br |
| web-ui | Change or review what a visitor sees on a site. Look first, plan with the design tokens, build, then test with worst-case data and screenshots. Use for each visible change, or when a page looks plain or generated. | webdesignreview |
| Skill | What it does | Tags |
|---|---|---|
| cientifico | Escreve e revisa texto científico ou técnico com fontes (artigo, proposta, relatório) em pt-BR, e em EN quando o texto é bilíngue. Use ao redigir, citar, checar números, preparar ou responder revisão. | writingpt-brresearch |
| easy-translate | Make an easy version or a controlled translation of a source document: easy pt-BR, Easy Japanese or controlled Simplified Chinese. Use for a versão fácil, a translated README or site string, or a stale translation. | writingtranslationpt-brjapanesechinese |
| essay | Check that an essay keeps the form that Montaigne gave the essay. The writer is in it, it opens on a case, it doubts and it does not teach. Use to check an essay, or to tell an essay from an article. | writingreview |
| ptbr | Reescreve texto técnico em português brasileiro por nível de rigor (estrito, descritivo, leve, científico, smart-note). Use ao escrever, reescrever ou revisar runbook, documentação, comunicado ou notas em pt-BR. | writingpt-br |
| ptbr-humanizer | Write or review the pt-BR text that a visitor reads on a site: UI strings, about pages and meta descriptions. The text must sound Brazilian, not like a translation or a model. Use for each change to visible pt-BR copy. | writingpt-brweb |
| ptbr-routing | Escolhe a skill e o nível de rigor antes de carregar uma skill de escrita em pt-BR. Use quando o pedido é escrever, reescrever ou revisar texto em português e não diz o nível nem a skill. | writingpt-br |
| ptbr-writer | Reescreve texto técnico em português brasileiro por nível de rigor (estrito, descritivo, leve, científico, smart-note). Use ao escrever, reescrever ou revisar runbook, documentação, comunicado ou notas em pt-BR. | writingpt-br |
| retorica | (em desenvolvimento) Retórica para texto técnico em português. Sem regras ainda: não use para reescrever texto. | writingpt-br |
| ste100 | Write or rewrite English text in ASD-STE100 Simplified Technical English. Use for docs, READMEs, PR bodies, tool descriptions, error messages, prompts and agent instructions, or when a text is dense, hedged or easy to misread. Not for pt-BR text (use ptbr-writer) or for creative copy. | writingenglishste100 |
| Skill | What it does | Tags |
|---|---|---|
| noaa-notices | List NOAA announcements new since the last snapshot that touch this repository (NWS Public Information Statements and Service Change Notices, NCEP model changes, NOAA-EMC/WW3 tags and production branches) and route each to the file it breaks or dates. Use when the user types /noaa-notices, asks what NOAA announced lately, whether GFS, GFS-Wave or NOMADS changed, or whether example 02's get_gfs.sh is still safe. | ww3noaaresearch |
| release | Cut a citable ww3-gpu release and keep its Zenodo/citation metadata right. Use when the user asks to release, tag a version, mint or update a DOI, or change CITATION.cff, .zenodo.json or the README citation badge. | ww3citationrelease |
| wfip | Write, revise or report on a Wave Forecaster Improvement Proposal (WFIP), the numbered design doc under docs/WFIPs/ that ties a non-trivial change to a deliverable of the project proposal and carries its definition of done. Use when the user says WFIP, WF-001, improvement proposal, propose, design before building, definition of done, DoD, Spec Kit or specify, or asks what the project has delivered against the proposal. | ww3planningdesign-doc |
| ww3-port | The WW3-specific traps of translating a WAVEWATCH III Fortran routine to C++/Kokkos and of claiming it regtested, learned on W3SNL1. Use while translating or reviewing a ported routine (after /port-next has scaffolded it), when a Kokkos result differs from the Fortran by a few ULP, when wiring a shim into W3SRCE or W3INIT, under OMPG/OMPH or ww3_multi, or before writing that a routine is 100% regtested. | ww3fortrankokkosport |
| ww3-port-rust | The traps of translating a WAVEWATCH III Fortran routine to no_std Rust for Rust-CUDA and rust-gpu. Use while translating or reviewing a routine, when a Rust result differs from the Fortran reference by a few ULP, or when choosing a float kind, a math call or a device buffer layout. | ww3fortranrustport |
| ww4-status | Check NOAA's WAVEWATCH IV (WW4) for changes since the last recorded snapshot and update this repo's research to match. Use when the user types /ww4-status, asks what WW4 has done lately, whether WW4 ported a source term or solver, adopted Kokkos, changed its spectral grid or released, or asks to refresh lesson 14 or the proposal's State of WW4. | ww3research |
Skills of other authors that these projects use. They are not in the plugin. common-skills gives the licence and the install command of each.
| Skill | What it does | Tags |
|---|---|---|
| ponytail | The simplest change that works, with review and audit skills | elsewherecodereview |
| frontend-design | A clear visual direction for a new or changed UI | elsewherewebdesign |
| webapp-testing | Tests a local web app with Playwright | elsewherewebtesting |
| design-taste-frontend | Landing pages and redesigns that do not look like a template | elsewherewebdesign |
| karpathy-guidelines | Rules against frequent LLM coding errors | elsewherecodereview |
| emil-design-eng | UI polish, component design and animation choices | elsewherewebdesign |
| review-animations | Reviews animation and motion code | elsewherewebdesignreview |
| break-ui | Tries to break a UI with worst-case data | elsewherewebtesting |
| mapbox | Seven Mapbox skills: cartography, styles, data, tokens, speed | elsewherewebmaps |
| humanizer | Removes the signs of AI writing from English text | elsewhereenglishwriting |
| humanizar | Removes the signs of AI writing from pt-BR text | elsewherept-brwriting |
| humanizer-pt-br | A different pt-BR humanizer | elsewherept-brwriting |