「the」の検索結果

Use when the user wants to know if something works — the answer requires running code, not analyzing it. Output is a verdict backed by evidence: passed, failed, or broken. Primary triggers: 'run the tests', 'does X still work after my change?', 'did the merge break anything?', 'verify the fix worked', 'check if the endpoint returns X', 'confirm nothing regressed', 'run tests/unit/test_foo.py', 'let me know the results', 'make sure my changes didn't break anything'. Hard stops — do NOT use for: reviewing test code for quality/coverage gaps, debugging why test infrastructure/databases/seed scripts are misbehaving, writing or fixing tests, diagnosing root causes of unexpected behavior. The deciding question: is the user asking for the result of executing something, or asking for help understanding/analyzing/improving something? If it's the latter, use diagnose or review-code instead.

日本語の概要は準備中です。原文の説明を表示しています。

Launch the kolu dev server on two RANDOM free ports so it never collides with the running production `kolu.service`, remember the ports for the rest of the session, and tear down only the dev instance. Load before running the app locally — for evidence capture, driving a live kolu with the chrome-devtools MCP, or any `just dev` you'd otherwise run by hand. Triggers on "run kolu locally", "launch the dev server", "boot kolu", "drive a live kolu", "start the app to screenshot it", or before pointing chrome-devtools at a local kolu.

日本語の概要は準備中です。原文の説明を表示しています。

Run a Meta-Harness-style optimization loop NATIVELY — automatically search over the scaffolding around a FIXED base model (memory, retrieval, context construction, prompt templates, summarization, tool-selection logic) by proposing candidate variants, scoring each on a cheap deterministic eval, and keeping a Pareto frontier of quality vs cost — using native Agent / Workflow / loop tools instead of a standalone Python harness. Use this whenever the user wants to optimize, evolve, tune, distill, or search over a harness, scaffold, prompt system, memory or retrieval policy, context-assembly code, or summarizer while keeping the model fixed; whenever they mention Meta-Harness, harness optimization, scaffold evolution, automatic prompt/memory optimization, an evolutionary or Pareto search over candidate implementations, or "make the harness/agent better without retraining"; and whenever the gain must come from the code AROUND the model rather than the model weights. Reproduces the Meta-Harness paper's method natively, with no claude_wrapper.py and no metered solver API.

日本語の概要は準備中です。原文の説明を表示しています。

The logical structure of mathematical proof -- how first-order logic, natural deduction, and semantic reasoning combine to produce rigorous mathematical arguments. Covers the logical skeleton of direct proof, contraposition, contradiction, cases, and induction; the role of definitions, lemmas, and theorems; how informal mathematical prose maps to formal deductive structure; and the difference between proof-as-object and proof-as-activity. Use when the question is "why is this a valid proof?" rather than "what proof technique applies?"

日本語の概要は準備中です。原文の説明を表示しています。

Deploys the current project to a live HTTPS URL via Cybrix. Activates on any request to make the current project public, get a URL for it, deploy it, ship it, host it, publish it, put it online, or make it live — including casual phrasings like "send this to the internet", "share this as a link", "give me a URL", "make this accessible online". Activates whether the user mentions Cybrix by name or not (when no other host like Vercel, Netlify, Cloudflare Pages, Railway, or Render is named). Also activates on equivalent phrases in other languages — Russian ("задеплой", "опубликуй", "выложи в сеть", "запушь это в инет", "сделай ссылку"), Spanish ("despliega esto", "publica esto", "ponlo en línea"), German ("deploy das mal", "stelle das online"), and other common languages where the intent is clearly to make a project publicly accessible via URL. Uses heuristic detection to determine if a project is static or needs a server runtime — works with any framework, not just a known list. Automatically detects and handles environment variables so the build receives everything it needs. Returns a live URL on *.cbrx.cc or a user-configured custom domain. Does NOT activate when the user explicitly names a different host (Vercel, Netlify, Cloudflare, Railway, Render).

日本語の概要は準備中です。原文の説明を表示しています。

Turns a theory into falsifiable, pre-registerable hypotheses, covering the DGP and DAG, identification, SATE versus PATE, counterfactual and directional framing, named estimands, justified SESOIs, the choice among NHST, equivalence, interval, and minimum-effect tests, scope conditions, and primary, secondary, and exploratory tiers. Use when the user asks to turn a research question or theory into hypotheses, whether a prediction is falsifiable, what the estimand is, or how to predict a null. Framing goes to narrative-building, the plan to pre-registration-writing.

日本語の概要は準備中です。原文の説明を表示しています。

Set up Claude Tag for oncall for a team, from that team's standing (public) oncall / monitoring channel. Two things: it sets the channel up for monitoring, so alerts and incident posts here (from a person or another Slack bot) get triaged and acted on automatically, and it finds and sets up the connectors, plugins and skills Claude uses for oncall. A workspace can have several such channels; run once in each, by anyone, and every run adds or updates that team's section in the one oncall memory in shared workspace memory. Use on "set up oncall", "init oncall", "configure Claude for on-call". Surveys the agent connectors an admin configured for Claude, proves one with a read-only pull so access is verified, explores tools, Slack, repos and docs to learn how oncall works here, then recommends what's missing and walks the team through setup. The result goes into the indexed oncall memory every channel reuses; the other skills (incident init, investigation, handoff) pick it up automatically.

日本語の概要は準備中です。原文の説明を表示しています。

Instruments and workflow for debugging the VirtualList — the chat transcript (`InfiniteList`) and the sidebar (`FiniteList`). Covers the built-in consistency checker, the on-screen overlay, the five `?vl*` URL flags, the touch-gesture rig, the frame recorder, attaching to a real Android device over CDP, and how to read a trace without being fooled by a measurement artifact. Use when the list jumps while scrolling, blanks, sticks on skeletons, or the overscroll/bounce misbehaves. The component's specification is `docs/ui/virtual-list.md`; this skill is only how to *measure* it. Pairs with /debug-ui (browser side) and /server-loop (server side).

日本語の概要は準備中です。原文の説明を表示しています。

Add an external probabilistic classifier to a decision path without letting it take the path over. Covers finding the externally-graded rows you are already logging, measuring the confidence separation between the overrides it got right and the ones it got wrong, choosing an operating point your data licenses, failing open at the call site, and proving the whole arrangement offline with no API key. Applies to any service returning a score you can order — TypeSafe AI's System One models (Jev) are the worked example. Most of the method transfers to a local model or a second heuristic, with the parts that do not called out where they arise. Use when a rule-based path is wrong often enough to hurt, when someone proposes replacing a heuristic with a model, when a confidence threshold needs a defensible value, or when an oracle already in production has never been graded.

日本語の概要は準備中です。原文の説明を表示しています。

Fix scattered/torn regions in hex-grid simulated annealing layouts that use "competitive flood-fill" (one tile per region per outer iteration). Use when: (1) regions of >20 tiles end up dispersed instead of forming compact blobs, (2) maxDist/sqrt(N) dispersion ratio is 10×+ worse than the ideal compact blob, (3) tiles assigned to one region show up in multiple disconnected fragments across the layout, (4) the visualization has a "trypophobia ring" pattern at the perimeter where some regions spill into a perimeter spiral, (5) an overflow spiral fallback exists in the layout code and frequently kicks in for large regions, (6) bigger regions look worse than smaller ones (small regions stay compact, big ones shred). Root cause: interleaved competitive flood-fill (round-robin one tile at a time across all regions) lets large regions get boxed in by neighbors before they finish growing — the unfilled remainder is dumped into the overflow spiral at the perimeter, far from the seed. Fix: replace with sequential region-by-region BFS — process regions in size-descending order, fully claim each region's N contiguous tiles before moving to the next.

日本語の概要は準備中です。原文の説明を表示しています。

"Apply Grounded Theory (Glaser and Strauss) to build theory inductively from qualitative data through open, axial, and selective coding. Use this skill when the user needs to develop new theory from data rather than test existing hypotheses, conduct theoretical sampling and constant comparison, determine when theoretical saturation is reached, or when they ask 'what theory explains this phenomenon', 'how do I code qualitative data systematically', or 'when do I stop collecting data'.".

日本語の概要は準備中です。原文の説明を表示しています。

Triage the Gaia issue backlog and skill-batch intake queue. Use this skill when someone asks to: "triage issues", "clean up the issue tracker", "review the backlog", "close stale issues", "process the intake queue", "review skill-batches", "evaluate draft skill proposals", "approve or reject pending skills", "is this issue still valid?", "what's clogging the backlog?", "run a triage pass", or /gaia-triage. Covers two workstreams: (1) GitHub issue lifecycle — identifying resolved, stale, or need-more-info issues and acting on them via gh CLI; (2) skill-batch intake — evaluating draft proposals in registry-for-review/skill-batches/ and routing them toward promotion or rejection. This is the gatekeeping step before new skills enter the canonical registry; and (3) a cross-repo triage sweep — sizing, P0-P4 prioritization, umbrella/epic synthesis, and assignment across every gaia-research org repo, on the one headquarters tracker.

日本語の概要は準備中です。原文の説明を表示しています。

Regenerate and commit Class S site-served artifacts after registry changes, then sync documentation across the main repo and the adjacent wiki. Use this skill when: you've changed registry/nodes/, registry/named/, or named-skills.json and need to regenerate docs/graph/* so the live website reflects the update; CI Guard E (docs-cohesion.yml) is failing because Class S artifacts are out of date; someone says "sync the docs", "regenerate docs", "run gaia dev docs", "update the site graph", "commit the Class S artifacts", "website is stale", or "update the wiki"; or you're finishing a registry PR and need to ensure docs/graph/gaia.json, docs/graph/gaia.gexf, docs/graph/gaia.svg, and docs/graph/named/index.json are committed alongside the source change. Also covers updating README.md, docs/index.html, and mirroring changes to the adjacent wiki repo at ../gaia-wiki.

日本語の概要は準備中です。原文の説明を表示しています。

Use when a task needs the judgment of a Compensation, Benefits, and Job Analysis Specialist — market-pricing a job against salary survey data, scoring a job through a point-factor or IPE job evaluation, running a pay-equity regression or cohort analysis, building or refreshing a salary structure and compa-ratio bands, classifying a job family's FLSA exemption status through the duties test, or modeling benefits plan cost (loss ratio, PEPM, self-funded vs. fully insured). Distinct from hr-people-manager (owns the comp/leveling judgment calls — whether to approve an exception, promote someone into a level), hr-specialist (owns individual-case leave/I-9/ADA/FLSA reclassification for one employee's drifted duties), and compensation-benefits-manager (owns pay philosophy, negotiation-range policy, and the department that runs this work). This role owns the mechanics: the survey match, the point score, the regression, the band math — the analysis those upstream policy and downstream judgment calls are made against.

日本語の概要は準備中です。原文の説明を表示しています。

Sync the yaah marketing website (website/index.html) with the current project state — features, CLI commands, skills, agents, hooks, and installation instructions. Use this skill whenever the user adds, removes, or changes a feature, CLI command, hook, skill, agent, MCP server, or LSP provider AND the website should reflect that change. Also use when the user explicitly asks to update, sync, or refresh the website, or says things like 'update the site', 'keep the website current', 'reflect this on the website', or 'the website is out of date'.

日本語の概要は準備中です。原文の説明を表示しています。

Run a Meta-Harness-style optimization loop NATIVELY — automatically search over the scaffolding around a FIXED base model (memory, retrieval, context construction, prompt templates, summarization, tool-selection logic) by proposing candidate variants, scoring each on a cheap deterministic eval, and keeping a Pareto frontier of quality vs cost — using native Agent / Workflow / loop tools instead of a standalone Python harness. Use this whenever the user wants to optimize, evolve, tune, distill, or search over a harness, scaffold, prompt system, memory or retrieval policy, context-assembly code, or summarizer while keeping the model fixed; whenever they mention Meta-Harness, harness optimization, scaffold evolution, automatic prompt/memory optimization, an evolutionary or Pareto search over candidate implementations, or "make the harness/agent better without retraining"; and whenever the gain must come from the code AROUND the model rather than the model weights. Reproduces the Meta-Harness paper's method natively, with no claude_wrapper.py and no metered solver API.

日本語の概要は準備中です。原文の説明を表示しています。

High Availability control plane for Yuzu — the canonical entry point for any work on ADR-2002 (active-active server tier, gateway-fronted routing, Postgres-backed coordination seam, HA Postgres storage). Use when the user says "/ha", "/high-availability", "/ha-matrix", asks to plan or implement an HA workstream (WS-0…WS-14), asks "what's our HA gap" or "is it safe to run a second replica yet", asks to audit HA readiness, or starts work touching the coordination seam (leader election / fenced epochs / transactional outbox), gateway intra-cluster routing, shared `AgentRegistry` presence, PKI/CA HA, or the `deploy/docker/*ha*` / `scripts/ha/` surface. The skill bundles the two-axis model, the safe-to-scale gate, the delivery matrix (current-state + reviewers + status), the per-workstream workflow, and the load order for the routed reference docs.

日本語の概要は準備中です。原文の説明を表示しています。

Run a task iteratively over a user-specified duration by dispatching subagents. The orchestrator stays strictly linear and time-checked, but each iteration fans out MULTIPLE subagents in parallel over independent units (and those subagents may fan out further). Shared state — a compact progress digest and a growing environment cheatsheet — lives on disk and is passed to every subagent BY REFERENCE, killing the per-subagent rediscovery tax. Every subagent first reads an initialiser preamble that points it at that shared state and makes it write findings back, so the cheatsheet populates itself; role prompts are produced ONCE on the filesystem by a single scaffold command and dispatched by path, never re-typed per dispatch. Subagents persist their results to disk and return only a tiny status, so the orchestrator's context stays small and lasts. The orchestrator classifies the goal as FINITE or OPEN-ENDED: finite lists may finish early; open-ended goals use the full duration, never idle, never manufacture busywork. Use when the user gives a task and a duration and wants it ground out iteratively by subagents over that time.

日本語の概要は準備中です。原文の説明を表示しています。

Catalog-then-fix slop removal on an existing prose draft. Parallel catalog subagents inventory four slop categories (theatrics, overwrought register, corrective antithesis, dramatic short sentences), fix subagents repair only what a catalog names, mechanical lint + frequency analysis bracket the run, and a regression gate (lint delta, word count, cold meaning check) decides whether the result ships. Use when the user says "deslop this draft", "catalog deslop", "strip the AI patterns from this file", or as the final stage of the article-factory and mosaic-writing skills. For writing or redrafting prose from scratch, use the prose skill instead - this skill only repairs an existing draft in place.

日本語の概要は準備中です。原文の説明を表示しています。

Render the GAME-DESIGN document (the GDD) into a human-readable form — read-only Markdown chapters under docs/design/ (index, systems, flows, content, economy, glossary), with facts resolved inline from GD-IDS.yaml and drafts flagged 🚧. Optional --web also renders an HTML site from the unikit-docs template (falls back to Markdown-only if the template is absent). Use to export or publish the design for people to read, e.g. "render the GDD", "export the game design to docs", "generate readable design docs", "publish the GDD", "make a design doc site". Read-only — it never edits the GDD or registry, only renders it. This renders the game DESIGN — for the project's CODE documentation (README, docs/*, modules, architecture) use /unikit-docs.

日本語の概要は準備中です。原文の説明を表示しています。

Curate .unikit/MCP-RECHECK-NOTES.md — the project's log of findings about the configured engine MCP server. Four jobs: re-stamp (the server moved, so every entry is suspect), replay (reproduce a `replay: safe` finding inside a disposable sandbox to see whether it still holds), retire (offer to drop what was fixed or went upstream), and upstream (print a ready diff for the packaged INDEX.md). Use for "audit the MCP notes", "are these findings still true", "recheck the MCP traps", "clean up MCP-RECHECK-NOTES", "the server was updated — revisit the notes". Replaying mutates a live editor, so it starts by telling you how to prepare it — save your scene, open an empty one — and asks once before touching anything. Everything it creates lives under a single UNIKIT_AUDIT_<runid> prefix, is deleted in one action, and the scene is never saved. To record a NEW finding use /unikit-mcp-trap instead.

日本語の概要は準備中です。原文の説明を表示しています。

Write, open and — when allowed — merge a pull request for the current feature branch of a {{engine_name}} project, with or without a plan. Builds a plain-language PR text (one feature per line) from the plan's `## Modules` and task `WHY:` lines, or — without a plan — from the commits and the code changes themselves; pushes the current state (asking where to end when it stands inside a plan module), opens or updates the branch's single open PR through the GitHub MCP, and merges with a merge commit only when the merge brings nothing from the base branch. Levels: remind (text only), create, merge; capped by `git.pull_requests.max_level`. Trigger phrases: "pull request", "open a PR", "create PR", "PR description", "merge the PR". For commit messages use /unikit-commit; for reviewing a PR's code use /unikit-review.

日本語の概要は準備中です。原文の説明を表示しています。

Implement a change end-to-end from a freeform prompt - settle the requirements, design the engineering approach and get it approved, implement with TDD, run the format/compile/credo/coveralls gates without shortcuts, self-review the diff like a senior engineer, QA the finished change in the running app, record a demo video when it is visible there, then commit and open the PR. Use when the user wants to "complete an issue", "work the ticket", or hands over a description of something to build.

日本語の概要は準備中です。原文の説明を表示しています。

Launch Marcus multi-agent experiments with independent CLI agents in tmux panes. Use when the user says "build X with Marcus", "run a Marcus experiment", "spawn N agents with Marcus", "/marcus", or references the Marcus multi-agent framework. This skill handles experiment setup, config generation, and spawning independent agent CLI processes that coordinate via the Marcus MCP server. Each agent runs in its own tmux pane with full autonomy. NOT for ClawTeam — this is specifically for the Marcus MCP-based multi-agent system. NOTE: "Marcus" is also the name of the MCP server that agents use for coordination. The /marcus skill launches experiments that USE the Marcus MCP server — they are complementary, not competing. The MCP server must be running before invoking this skill. The Marcus MCP server is agent-agnostic; this skill supports Anthropic's claude CLI, OpenAI's codex CLI, and Google's gemini CLI as agent harnesses (selected via --harness).

日本語の概要は準備中です。原文の説明を表示しています。