ワンクリックで
overdeck
overdeck には eltmon から収集した 115 個の skills があり、リポジトリ単位の職業カバレッジとサイト内 skill 詳細ページを表示します。
このリポジトリの skills
Add repositories to the current progressive polyrepo workspace
pan close <id> — close-out ceremony for a completed and merged issue
pan review <subcommand> — manage the code review lifecycle: list pending work, re-request review, heal status drift, reset/abort/restart review cycles
pan start <id> — spawn a work agent for an issue in its own tmux session and workspace
Maintain a project knowledge wiki in Open Knowledge Format with /okf init, open, author, convert, sync, study, retro, extract, validate, lint, and embed.
pan flywheel — start, pause, resume, complete, stop, inspect, emit, and report on the singleton Fix-All Flywheel orchestrator
pan orders — create, inspect, edit, and start Flywheel order books
pan knowledge <id> — spawn a knowledge agent to maintain the project's Open Knowledge Format (OKF) bundle
Create Overdeck repo commits that satisfy commitlint and husky on the first try
Test the full specialist handoff pipeline (review → test → merge)
Standard workflow for implementing new features with testing
pan admin config <cmd> — view and edit Overdeck project configuration
pan approve has been removed — use the dashboard MERGE button instead
pan doctor [options] — check Overdeck system health, dependencies, and configuration
Complete setup for registering a new project with Overdeck. Handles project registration, issue prefix, workspace config, trust setup, xBRIEF task support init, tracker config, and validates against working projects.
Test the Overdeck framework by supervising an agent through the full lifecycle, identifying and filing every bug encountered
pan plan <id> — start issue planning, including non-interactive --auto mode; also finalize/done planning artifacts
Quick start guide combining installation, setup, and first workspace
Executive recap of the most recent day's work (default last 24h) across the Overdeck pipeline — what shipped, what's still in flight, what needs the operator's attention, and the numbers — PLUS a separate, shareable user-facing PR-FAQ of what was released. EVERY issue is described in plain language (what it is / what it does), never a bare ID. Use when the user asks "what happened overnight / today / while I was away", "catch me up", "daily recap", "standup", "what did the flywheel do", or wants an up-leveled summary of recent activity.
pan <verb> <args> — umbrella dispatch for all Overdeck CLI commands. Invoke bare to see the six-bucket taxonomy, or pass a full command to run it.
Start and recover Overdeck swarm slots for planned issues with `pan swarm`; use for parallel work dispatch or failed-slot recovery.
pan sync — distribute skills and agents to every supported Overdeck harness
pan task <verb> <issue> — claim and complete xBRIEF checklist items
pan wipe <id> — destructive reset to Todo: remove workspace, processes, branches, review state, xBRIEF task state, and tracker status
List the issues in the Overdeck pipeline, grouped by phase (Ship · Review · Work · Plan) with their TITLES — the on-demand text form of what the dashboard shows. Two views: the default "resource" view mirrors the Command Deck project tree (issues with a live workspace/branch/agent/PR right now); `--phase` gives the "lifecycle" view that mirrors the Pipeline page (every issue by phase, including planned-but-not-started work). Use when the user asks "what's in the pipeline", "list the pipeline", "what are we working on", or wants the active issues by phase with names. For the dense per-issue status matrix use /pipeline-status; this is the grouped, titled list.
Cross-room visual status board for every active issue moving through the Overdeck pipeline. One row per issue, one column per phase (agent → review → test → verify → merge → ready), with a checkmark or X in each cell. Designed to be readable from across the room while agents work autonomously. Surface this BEFORE the verbose pan-status / agent-status output whenever the user asks "where are we?" or wants a status overview.
Opus-driven planning for issues before Sonnet implementation. Creates workspace, .pan/continue.json, .pan/spec.vbrief.json with xBRIEF tasks, and updates issue tracker. Ensures strategic decisions are made by Opus, not cheaper models.
Compatibility redirect: /write-vbrief has been renamed to /write-xbrief
Write an xBRIEF spec and continue.json directly — without launching the interactive planning agent. The workspace working copy deliberately remains .pan/spec.vbrief.json for legacy compatibility. Use when the work is well-understood and the agent can author the plan from the issue body and codebase alone. Also use when you ARE the work agent and need to self-plan before implementing. Covers the full xBRIEF v0.8 schema, continue.json format, task sizing rules, inspection gates, and the `pan plan finalize` handoff command.
pan restart — scoped restart (dashboard by default; --cliproxy, --traefik, or --full) that will not strand shared sidecars
pan up — start the Overdeck dashboard (Node 22, port 3011)
Build Overdeck, then restart the dashboard only if the build succeeds.
pan project <subcommand> — add, rename, remove, and manage Overdeck-monitored projects
Install an external skill or Claude Code plugin into Overdeck's bundled distribution so pan sync ships it to every Overdeck user
Verify Overdeck agents are ACTUALLY progressing by READING each agent's real output and assessing what it is doing — not just whether a session is alive or a pane changed
Deep performance analysis focusing on algorithms and resources
pan handoff <conv> — agent-authored conversation handoff that spawns a new conversation. The focus text MUST be ≤500 characters or the fork is rejected.
pan hygiene [options] — audit whether orchestration work is committed, pushed, mergeable, live, and clean
pan rollout <status|retry> <id> — inspect or re-run post-merge release coordination
Safely restart the Overdeck dashboard server (production Node 22 dist) using a detached process. Use when pan up / pan restart hang, the dashboard churns/reconnects in a loop, two dashboards are dueling, or the server is running from a workspace path / wrong port.