Menu
Skills dans ce dépôt
SkillsMP a collecté 447 skills depuis jleechanorg/claude-commands. Ouvrez un skill pour examiner sa source et ses détails.
jleechanorg/claude-commandsAffichage de 40 skills collectés sur 447.
RETIRED — self-hosted-oss/install.sh and its runtime mirror sync flow no longer exist. Historical reference only; see runner-health and ezgha-watchdog for the current mechanism.
Langue du texte source : anglais
Pre-flight checklist for self-hosted runner failures. Use when investigating low disk, missing runner, or stuck Green Gate on the $GITHUB_REPOSITORY fleet. Always verifies host-level container state, not just GitHub API.
Langue du texte source : anglais
Spawn or respawn a persistent, crash-recoverable Sonnet sidekick for ANY long-running mission — PR/CI fleets, research sweeps, migrations, monitoring, ops runbooks, data pipelines, multi-day projects in any repo or no repo at all. The sidekick is ALWAYS a…
Langue du texte source : anglais
Define and run skeptic exit criteria for non-trivial tasks — independent verification agent with inverted incentive to find gaps
Langue du texte source : anglais
Turn any feature, script, or workflow into a properly-skilled, tested, auditable Hermes/Claude skill. Use when the user says skillify, is this a skill, make this proper, or add tests and evals. Runs the 10-item skillify checklist against the target and…
Langue du texte source : anglais
Audit Slack channel threads for untracked work items — checks each thread against GH issues/PRs, reports gaps, and optionally redrives dropped threads by posting a status reply to the channel. Slash command: /slack-audit.
Langue du texte source : anglais
Run /smoke-local — boot the real local MCP backend through the test harness, route all LLM calls through the agy CLI provider (never mock, never direct Gemini), play a new campaign to character level 4, and probe combat/spell/inventory/NPC-dialog narrative…
Langue du texte source : anglais
Orchestrate multi-agent swarms (ultracode Workflow fan-outs + Agent-team lanes) with adversarial verification, /innov, /advice, and a sidekick durability layer for crash-recoverable runs
Langue du texte source : anglais
Use when executing implementation plans with independent tasks in the current session
Langue du texte source : anglais
Use when starting any conversation - establishes how to find and use skills, requiring skill invocation before ANY response including clarifying questions
Langue du texte source : anglais
Use when you have a spec or requirements for a multi-step task, before touching code
Langue du texte source : anglais
Use when creating new skills, editing existing skills, or verifying skills work before deployment
Langue du texte source : anglais
When asked to verify whether a Claude Code feature works (especially slash commands, dialogs, pickers, status indicators), spawn a real interactive TUI session in cmux — never use `claude --print "/feature"` as a test, because --print is non-interactive and…
Langue du texte source : anglais
Record captioned browser/UI evidence videos, upload to GitHub PR attachments, and link sanitized artifacts in a self-contained gist
Langue du texte source : anglais
Use when the user invokes /up or asks to persist a coding-agent rule or preference. Default: repo-level Markdown under the active repo. Use /up --repo for any repo, /up --global for the cross-runtime surfaces.
Langue du texte source : anglais
Produce zero-code, zero-tech product specs and VISUAL user stories for any project/product — what the user sees, hears, and does, at a bar where a stranger could rebuild the experience without reading code. Slash command /user-story.
Langue du texte source : anglais
your-project.com-specific practice for the no-code visual user-story spec — where the docset lives, which evidence sources settle which claim, the traps that have produced false claims here, and the claims already refuted. Use with (not instead of) the…
Langue du texte source : anglais
Use when asked to analyze a video file and add visual description captions or text overlays burned into the output video
Langue du texte source : anglais
Add burned-in captions to a screen-recording / gameplay video WITHOUT blocking the video content. Use when the user wants captions ABOVE the gameplay (e.g. "put captions above", "don't block the video", "letterbox the captions", "extend the video vertically…
Langue du texte source : anglais
Browser-based multi-model review using ChatGPT, Gemini, Grok, and Perplexity Web via aside-mcp. Use when you need an independent multi-model adversarial pass over a PR, evidence bundle, or design doc — not for in-session reviews (use /advice for that).
Langue du texte source : anglais
Assess wiki structure and ratios against Karpathy pattern standards
Langue du texte source : anglais
Write an outcome-driven goal spec that drives agents toward verifiable outcomes — solid evidence, green-CI PRs, high code quality. Pulls context from the last 30 days of coding-CLI convos (/history) and cross-session memory (/ms) so the goal is grounded in…
Langue du texte source : anglais
Use this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting,…
Langue du texte source : anglais
Use when designing or reviewing model-owned contracts, output schemas, backend guards, routing, semantic parsing, intent detection, or data fields where application code might duplicate judgment or ask a model for redundant facts.
Langue du texte source : anglais
Canonical definition of zero-touch-by-operator for PR autonomy measurement. Use when evaluating, reporting, or labeling PR autonomy status.
Langue du texte source : anglais
Use when designing or reviewing backend adjusters, corrections, defaults, suppressions, clamps, guardrails, compatibility aliases, or fallback logic that may override or reshape model-owned output.
Langue du texte source : anglais
ZFC Level-Up architecture alignment — file boundaries, API contracts, pre-flight checks, and milestone status. Read this BEFORE any level-up/rewards/XP code change.
Langue du texte source : anglais
为 Hermes 配置 MiniMax 作为模型源。MiniMax 提供两种接入方式：API Key 直连（openai-completions 协议）和 OAuth 门户（anthropic-messages 协议）。包含 provider 注册、模型定义、别名配置、fallback 链接入和验证的完整流程。当管理员说想"加 MiniMax"、"配 minimax"、"接入 MiniMax 模型"、"加海螺模型"、"配 M2.1"时使用此 skill。
Langue du texte source : chinois
Hermes wrapper for the `/advice` token-efficient second-opinion slash command. Canonical source is `~/.claude/skills/advice/SKILL.md`. Adds docs-accuracy review template, patch-review template, and three operator-correction pitfalls verified 2026-07-30: #14…
Langue du texte source : anglais
Build agent memory patterns that prevent the same mistakes — using session replay, error tracking, and a self-correcting knowledge base.
Langue du texte source : anglais
Use when installing any agent-development CLI (GSD Core, Superpowers plugin, mattpocock/skills via npx, marketplace plugins, npx/npm/pip wrappers) into Claude Code or a sibling runtime. External installers silently write always-on hooks to…
Langue du texte source : anglais
Get a status report of all PRs agento is handling — merged vs not merged, green status breakdown.
Langue du texte source : anglais
Delegate coding change tasks to Agent-Orchestrator (AO). Triggered by the keyword "agento" anywhere in a message, or by default for any coding task. Never use mctrl unless explicitly requested. Default worker is `MiniMax-M2.7` via `agent-minimax` plugin — do…
Langue du texte source : anglais
Spawn an Agent-Orchestrator (AO) worker that uses the minimax CLI with the M3 model (via the minimax Anthropic-compatible API). Use when: the task should be dispatched to a worker that uses minimax M3 specifically, when the user says 'use minimax', 'use the…
Langue du texte source : anglais
Diagnose "is AO actually working?" by cross-referencing live OS processes, the AO database (`~/.ao/data/ao.db`), and GitHub PR commit history — bypassing the `ao-progress-reporter.sh` and MCP Agent Mail daily report, both of which can show stale or fabricated…
Langue du texte source : anglais
"Use when the user asks for a multi-file edit + push + PR task (>5 turns) and steers away from AO worker dispatch. Print mode `claudem -p` is too short for multi-step tasks; tmux mode lets the worker drive to durable end-state while the gateway session polls…
Langue du texte source : anglais
Aside browser is the PRIMARY browser tool for all Hermes/Claude/Codex work. Use aside CLI (repl/mcp/exec modes) or aside-mcp before Playwright MCP / superpowers-chrome. Use when opening browsers, scraping, UI tests, localhost verification, cookie injection…
Langue du texte source : anglais
Pitfalls when using `codex exec` to drive a third-party CLI tool, skill, plugin, or workflow (NOT for using Codex as a coding agent — see the bundled `codex` skill for that). Class-level skill. Triggers when the task involves driving any external tool via…
Langue du texte source : anglais
Detect enabled babysit crons whose referenced PR is MERGED or CLOSED and disable them. Companion to the in-script `is_pr_terminal()` check in babysit.py. Without this watchdog, babysit crons can run hundreds of polls for weeks after the PR they were guarding…
Langue du texte source : anglais
Force-purge a launchd/cron-driven home-config backup folder (backup/, snapshot/, home-config/) from a public GitHub repo, then prevent recurrence. Trigger when (a) `git push` keeps re-pushing personal home-dir content to a public repo every N hours, (b) `gh…
Langue du texte source : anglais