Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
직접 명령은 검토 Prompt를 거치지 않습니다. 실행하기 전에 소스를 확인하세요.
npx skills add https://github.com/majiayu000/claude-skill-registry --skill quickstart명령은 한 줄로 유지됩니다. 복사하기 전에 가로로 스크롤해 전체 내용을 확인하세요.
로컬 사본을 원하시나요? SkillsMP에서 현재 제공할 수 있는 파일을 다운로드하세요.
LLM token logprobs and calibration. Per-decision confidence, ECE, Brier, reliability diagrams, low-confidence triage.
Analyze LLM token logprobs and calibration. Use for per-decision confidence, ECE, Brier scores, reliability diagrams, and low-confidence triage.
回顾最近 N 天的 Claude Code 使用记录——扫描原始会话数据,按主题分组汇总"我都做了什么",并从个人操作系统视角输出模式、风险与增删建议。当用户说 /recap、"看看我这几天做了什么"、"回顾一下我最近的会话"、"这两天我用 claude 干了啥"、"活动回顾" 时使用。
SKILL.md 표시 중
SOC 직업 분류 기준
| name | quickstart |
| description | Show AgentOps next action. |
| practices | ["pragmatic-programmer","agile-manifesto"] |
| hexagonal_role | driving-adapter |
| consumes | ["rpi"] |
| produces | ["stdout"] |
| context_rel | [{"kind":"customer-of","with":"rpi"}] |
| skill_api_version | 1 |
| context | {"window":"inherit","intent":{"mode":"none"},"intel_scope":"none"} |
| metadata | {"tier":"session","dependencies":[]} |
| output_contract | stdout: onboarding walkthrough |
One job: Tell a new user what AgentOps does and what to do first. Fast.
YOU MUST EXECUTE THIS WORKFLOW. Do not just describe it.
git rev-parse --is-inside-work-tree >/dev/null 2>&1 && echo "GIT=true" || echo "GIT=false"
command -v ao >/dev/null && echo "AO=true" || echo "AO=false"
command -v bd >/dev/null && echo "BD=true" || echo "BD=false"
[ -d .agents ] && echo "AGENTS=true" || echo "AGENTS=false"
[ -d "$HOME/.agents" ] && echo "GLOBAL_AGENTS=true" || echo "GLOBAL_AGENTS=false"
[ -n "${CODEX_THREAD_ID:-}" ] || [ "${CODEX_INTERNAL_ORIGINATOR_OVERRIDE:-}" = "Codex Desktop" ] && echo "CODEX=true" || echo "CODEX=false"
Output exactly this (no additions, no diagrams):
AgentOps is the operational layer for coding agents.
It gives your coding agent four things it doesn't have by default:
Bookkeeping — sessions capture learnings, findings, and reusable context in .agents/
Validation — /council, /pre-mortem, and /vibe challenge plans and code before shipping
Primitives — skills, execution packets, and the ao CLI give you building blocks for almost any interaction
Flows — /research, /implement, /validation, and /rpi can run alone or compose end to end
Key skills: /rpi /research /validation /implement /council /pre-mortem /swarm /status
Full reference: /quickstart --catalog
Match the first row that applies. Output that message.
| Condition | Message |
|---|---|
| GIT=false + AO=true + GLOBAL_AGENTS=true | "🗂 You're outside a git repo but have a global corpus at ~/.agents. Global knowledge workflow:\n 1. /harvest — scan all .agents/ across your repos and promote artifacts into ~/.agents/learnings/\n 2. /compile — mine, synthesize, and write an interlinked wiki into .agents/compiled/ (runs from cwd; set AGENTOPS_COMPILE_RUNTIME=claude-cli if you have the claude CLI, no API key needed)\n 3. /knowledge-activation — turn the compiled corpus into playbooks, a belief book, and runtime briefings for future sessions\n 4. /status — flywheel health snapshot\nIf instead you want to start a fresh repo-local project here, git init first." |
| GIT=false | "⚠ Not in a git repo. Run git init first.\n (If you meant to work against your global ~/.agents corpus, run /quickstart from a dir with .agents/ or see /harvest, /compile, /knowledge-activation.)" |
| AO=false + CODEX=true | "📦 Install ao CLI first:\n brew tap boshu2/agentops https://github.com/boshu2/homebrew-agentops\n brew install agentops\n ao quick-start\nThen: /rpi \"a small goal\" to run your first cycle.\nUse /bootstrap after the core seed when you want PRODUCT.md, README.md, and PROGRAM.md." |
| AO=false | "📦 Install ao CLI first:\n brew tap boshu2/agentops https://github.com/boshu2/homebrew-agentops\n brew install agentops\n ao quick-start\nThen: /rpi \"a small goal\" to run your first cycle.\nUse /bootstrap after the core seed when you want PRODUCT.md, README.md, and PROGRAM.md." |
| AGENTS=false + CODEX=true | "🌱 ao is installed but not initialized here.\n Run ao quick-start to apply the repeatable core seed. ao quickstart is the stable alias.\n Then run /bootstrap only if you want the product/operations layer: PRODUCT.md, README.md, PROGRAM.md/AUTODEV.md, and optional hooks.\nThen: /rpi \"a small goal\" to run your first cycle." |
| AGENTS=false | "🌱 ao is installed but not initialized here.\n Run ao quick-start to apply the repeatable core seed. ao quickstart is the stable alias.\n Then run /bootstrap only if you want the product/operations layer: PRODUCT.md, README.md, PROGRAM.md/AUTODEV.md, and optional hooks.\nThen: /rpi \"a small goal\" to run your first cycle." |
| BD=false + CODEX=true | "✅ Codex plugin path ready.\n /rpi \"your goal\" — full /discovery → /crank → /validation pipeline\n /research <topic> — explore the codebase\n Default installs are hookless; native hooks are optional with install-codex.sh --with-hooks.\n Legacy explicit fallback commands remain ao codex ensure-start / ao codex ensure-stop when needed.\n Manual escape hatch: ao codex status\n Want issue tracking? brew install boshu2/agentops/beads && bd init --prefix <prefix>" |
| BD=false | "✅ Flywheel active. Start now:\n /rpi \"your goal\" — full /discovery → /crank → /validation pipeline\n /validation — close out recent work and capture learnings\n /research <topic> — explore the codebase\n Want issue tracking? brew install boshu2/agentops/beads && bd init --prefix <prefix>" |
| BD=true + CODEX=true | "✅ Codex full stack ready.\n ao reconcile --json — check git, CI, release, beads, and .agents agree\n bd ready — see open work after high-severity reconciliation findings are clear\n /rpi \"your goal\" — start a new goal from scratch\n Default installs are hookless; native hooks are optional with install-codex.sh --with-hooks.\n Legacy explicit fallback commands remain ao codex ensure-start / ao codex ensure-stop when needed.\n Manual escape hatch: ao codex status" |
| BD=true | "✅ Full stack ready.\n ao reconcile --json — check git, CI, release, beads, and .agents agree\n bd ready — see open work after high-severity reconciliation findings are clear\n /rpi \"your goal\" — start a new goal from scratch\n /status — see current session state" |
Starting a new project? Run /scaffold <language> <name> to generate project structure with best practices.
User says: /quickstart
What happens: Agent detects tools, shows one-line status, gives the single next action to run.
User says: /quickstart
What happens: Agent detects full stack is ready and suggests ao reconcile --json, bd ready, or /rpi "your goal".
| Problem | Solution |
|---|---|
| Skills not installed | bash <(curl -fsSL https://raw.githubusercontent.com/boshu2/agentops/main/scripts/install.sh) |
| Hook activation | AgentOps 3.0 is hookless — there is no ao command or flag that installs hooks, and CI is the authoritative gate. Hooks are opt-in and author-it-yourself via the hooks-authoring skill. |
| Flywheel count is 0 | First session — run /rpi "a small goal" to start it |
| Want the full skill catalog | Ask: "show me all the skills" or see references/full-catalog.md |
references/quickstart.feature — Executable spec: detect setup (git/ao/bd/.agents), recommend a state-appropriate next action (soc-qk4b)