注释

Ajisai — Agent Writing Protocol (SKILL.md)

masamoto1982

code-explainer

saajunaid

Explain code with visual diagrams, analogies, and step-by-step walkthroughs. Use when explaining how code works, teaching about a codebase, or answering "how does this work?"

3d-cv-labeling-2026

slurpyb

Expert in 3D computer vision labeling tools, workflows, and AI-assisted annotation for LiDAR, point clouds, and sensor fusion. Covers SAM4D/Point-SAM, human-in-the-loop architectures, and vertical-specific training strategies. Activate on '3D labeling', 'point cloud annotation', 'LiDAR labeling', 'SAM 3D', 'SAM4D', 'sensor fusion annotation', '3D bounding box', 'semantic segmentation point cloud'. NOT for 2D image labeling (use clip-aware-embeddings), general ML training (use ml-engineer), video annotation without 3D (use computer-vision-pipeline), or VLM prompt engineering (use prompt-engineer).

github-attachment-upload

handlename

Upload an arbitrary local file to GitHub user-attachments by driving a real browser with playwright-cli against an issue editor, then return the attachment URL. File-type agnostic — works for SVG, PNG, screenshots, or any file GitHub's editor accepts; it does not create an issue or comment. Designed to be called from other skills (e.g. plan-to-issue, create-pr) through a subagent whose only return value is the URL on success or a failure reason on failure. Use when a local file must become a GitHub user-attachments URL. Triggers include "このファイルをGitHubのuser-attachmentsに上げてURLをちょうだい", "upload this file to GitHub user-attachments", "get me an attachment URL for this image".

arinhub-submit-code-review

arinhubcom

Use this skill to submit code review when using the "ah" prefix. Use when asked to "ah submit code review 123", or "ah submit code review to PR 123". Submit a completed code review with line-specific comments and suggestions to a GitHub PR.

slot-claim

surreptakos

Claim a scheduled routine's Slot on its Claim register before doing any work, recheck before each write, mark each write, and post done. Use at the start of every desktop routine and Fallback Routine run, whenever a routine prompt names slot-claim, and to read a register's status.

Visual Storyteller

anorbert-cmyk

Creates narrative-driven visuals to explain value and increase conversion, optimizing for clarity and trust.

rust-error-handling

MILTONADINA

Rust error-handling patterns and anti-patterns -- unwrap() in request hot paths, panic! instead of returning a Result, leaking error internals via Debug derive on error types containing privileged state. Patterns covered include thiserror for typed errors, anyhow for application-layer error propagation, and error-payload sanitization at API boundaries. Triggers on Rust error-prone code authoring (unwrap, expect, panic!, Result<, > patterns). General security hygiene at the language-runtime layer.

code-review

olivoil

Review a PR and post findings as inline comments to GitHub. Shows all feedback for approval before posting.

curate-backlog

hoopit

Clean up the agent tracker board — duplicates, merges, stale and invalid issues, missing triage.

kata-intent

walternagai

Fase INTENT do ciclo Karpathy (kata). Use quando o agente @kata estiver na fase 2.5 — verificar intenção entre código, teste e especificação antes de mudar comportamento. Triggers: INTENT, intent gate, intenção, código vs teste, spec, conflito, fable-method.

sofa-v3

blinds123

Builds a complete DTC landing page using mechanism-first angle discovery, 4-specialist council copy, ENGAGE/FIBS/3-Secrets frameworks, vault-anchored writing, and the real anchor-template build system.

academic-orchestrator

ShijianRuan

Multi-phase academic research and writing orchestrator. Chains deep-research, academic-researcher, medical-imaging-review, citation-management, latex-paper-en, fact-check, and peer-review into a single quality-gated pipeline. Use when the user wants to produce a full academic paper, survey, or literature review with verified claims, proper citations, and submission-ready output. Triggers on: "write a paper", "produce a survey", "academic review", "write a review paper", "publishable review", "systematic review with verification", or explicit requests to use the orchestrator.

tc-fe-prep-workflow

Thitic9203

Prepare frontend manual test cases from a Jira story (AC/EC) with mandatory AC/EC coverage review and ISTQB/29119-3 quality check, then draft table in chat, export CSV/Excel, publish one comment on that story only, and close with a four-axis final TC review report (AC/EC alignment, spelling, numbering, scope). Use when the user asks for FE test cases, manual TC from acceptance criteria, draft TC comment on Jira, or TC FE Preparation from Helix (/tc-fe-prep). Do NOT use for API-only Swagger test cases (tc-api-prep-workflow), Playwright execution (testing-ticket-workflow), retest-after-fix (retest-bug-workflow), or opening bug tickets (create-bug-workflow).

cowart-image-edit

hwy-game

Revise a Cowart canvas image from an annotation screenshot using Vetta edit_image, then place the result with Cowart MCP. Use when the user submits a Cowart annotation edit or asks to clean up an annotated screenshot into a revised bitmap beside the original.

loom-guide

rjwalters

"Keeps the backlog healthy: tier labels, orphan verification, unblocking, epic tracking, living docs"

guardian

nitedani

"Guardian rounds on a PR (bloat and quality, part of Loop B in converge): the LeanKeeper charter, the guardian and implementer briefs, landing, and when a scope has converged."

en

zheref

Take one open pull request from the moment it opens to verified readiness and its merge (the run's own PR, at its terminus), under a mandatory acting-cycle cap — the landing report, fresh current-head observation, Sharingan remediation, catch-up when behind, and the bell only at Ready. Use when /mukai hands over its PR, or when the maintainer invokes /en on #<N> for a PR that is already open. Pending CI or review keeps the run active; quiet polls spend no cycle. En never merges and never casts a review vote. After this run completes, the regular pipeline's next phase is /third-hand — En does not own it.

mugetsu

zheref

Publish one release to one production destination — a store, the live environment — by running the repository's declared release row (nen shu release) and/or nen shu deploy --target <production target> --run, after nen release preflight reads green and the tag has been cut. Use ONLY when the maintainer records a per-target go in their own words; that message is quoted verbatim in the report and on the release PR, and is never inferred. This is G3 (CON-6): no composite reaches it on its own authority — not getsuga, not en, and futon only as a typed then step carrying the ADVANCE GO its § 3 defines (the maintainer's own invocation, bound to the tag that run cuts, one target, once, and only on a repository nen repo classify reports as kind process or library) — no agent proposes it, a go with no tag is refused unless it is that advance go, and one go publishes one target once.

ask-shuff

patrickshuff

Ask Patrick Shuff a question and get a response in his voice, drawing on his philosophy and perspectives on technology, startups, and engineering

local-dev

alamine42

Launch the app locally in a dev playground with portless and Agentation so the user can annotate.

explain-code

avallarino-ar

Explains code with visual diagrams and analogies. Use when explaining how code works, teaching about a codebase, or when the user asks "how does this work?"

email-swipe-file

imsurajj

100+ annotated real-world email templates for startups and developers. Use when the user asks to write cold emails, onboarding sequences, sales follow-ups, investor updates, launch announcements, re-engagement emails, transactional emails, or any startup email copy. Includes subject line formulas and copywriting rules.

sql

saajunaid

Write high-quality, optimized SQL with best practices for performance, NULL handling, security, and readability. Database-agnostic patterns with dialect-specific notes.