kickoff
Skill7 starsChanged 6 months agoMITView on GitHub
Kick off GoC in a fresh repo — introduce GoC, pick a persona, confirm the AGENTS.md merge, scaffold via goc install. AUTO-INVOKE on "kickoff", "use GoC here", "set up game of cards", or when any GoC skill runs in a repo with no deck directory.
36 other projects use a version of this.
Adopted by
Most recently taken up by
- Lucface/claude-skillstheir own version, first seen Nov 16, 2026
- arthuroc21/claude-skillstheir own version, first seen Nov 16, 2026
- krystaleicher/claude-skillstheir own version, first seen Nov 16, 2026
- OhmsSweetOhms/claude-skillstheir own version, first seen Nov 16, 2026
- fullstackhouse/skillstheir own version, first seen Nov 15, 2026
- elfrost/ai-patchlabtheir own version, first seen Nov 9, 2026
Used with
Projects with it also use grill, prompt-engineer and 6 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
renatorozasSkillgrill
Adversarial code review. Don't let me ship until the changes pass your scrutiny.
Also in 4 projects that use this
JeffallanSkillprompt-engineer
Writes, refactors, and evaluates prompts for LLMs — generating optimized prompt templates, structured output schemas, evaluation rubrics, and test suites. Use when designing prompts for new LLM applications, refactoring existing prompts for better accuracy or token efficiency, implementing chain-of-thought or few-shot learning, creating system prompts with personas and guardrails, building JSON/function-calling schemas, or developing prompt evaluation frameworks to measure and improve model performance.
Also in 3 projects that use this
MicrosoftSkillhandoff
Compact the current conversation into a handoff document for another agent to pick up.
Also in 7 projects that use this
AnthropicSkillstatus
Case status summary by audience — client-facing (plain language), internal (for the professor), or court-ready (formal caption format per local rules). Same facts, different framing and depth. Use when a student needs to update the client, brief the professor, or prepare a court status report.
Also in 3 projects that use this
SentrySkillreview
Fixture review skill.
Also in 5 projects that use this
OpenAISkillwriting-plans
Use when you have a spec or requirements for a multi-step task, before touching code
Also in 3 projects that use this
Other versions36 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 36
- Maximepodgorski/agent-skills★ 21same name
- syutolee/digital-marketing-skills-pack★ 10same name
- ai-analyst-lab/interview-prep-with-ai★ 9same name
- 0x43e96f/skills★ 8same name
- takleb3rry/claude-global-appbuilding-skills★ 6same name
- ArangoGutierrez/claude-toolkit★ 4same name
- adrian-wulf/vorarbeiter-code★ 2same name
- bluewhalelabs/claude-skills★ 2same name
- elfrost/ai-patchlab★ 2same name
- Kohulan/orthonym-skills★ 2same name
26 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository19 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- site/llms.txtllms.txt
- advance-cardSkill
- audit-deckSkill
- card-schemaSkill
- claude-kickoffSkill
- create-cardSkill
- decide-cardSkill
- deckSkill
- finish-cardSkill
- next-cardSkill
- pull-cardSkill
- refine-deckSkill
- retrospectiveSkill
- scan-deckSkill
- standupSkill
- tune-cadenceSkill
- upgradeSkill
The file
We split this skill into the rules it gives an agent
3 rules. 0 are common, 1 is distinctive, and 2 are very specific to this project.
0
Common
Most of the 29 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 35
Read all of the on-disk signals at once.
- Very specific: line 63
Do not proceed without goc.
- Very specific: line 180
After goc install returns, verify .game-of-cards/deck/ exists and the chosen file contains a <!-- BEGIN GOC --> block before continuing.
“Common” is measured: kept by at least 6 in 10 of the 29 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
When to invoke
Invoke when the user says "kickoff", "use GoC here", "set up game of cards", "initialize GoC", or when any GoC skill is first used in a repo with no .game-of-cards/deck/ directory. Host-agnostic: per-host complements (claude-kickoff, codex-kickoff, openclaw-kickoff) handle host-specific UX.
Kick off GoC in this repo
This skill is the host-agnostic onboarding dialog: it introduces GoC,
asks a few focused questions, then runs goc install to scaffold project
state. It runs idempotently — every stage detects on-disk state before
asking, so a re-run on a partially set-up repo only asks for the answers
it cannot already derive.
The body covers what every host needs. Host-specific UX (permission
prompts, plugin install cadence, private-notes file) lives in a
complementary per-host skill that runs after this one — for example,
claude-kickoff on Claude Code.
Details live in reference.md — a sibling file in this skill's
directory. Read the named section only when the situation applies:
| Situation | reference.md section |
| The user asks what install will write / wrote | What gets installed |
| Stage 3 persona is Classical team | Team tip: external deck location |
| The repo uses git worktrees | Worktrees |
| Several humans + agents share one deck | Multi-team coordination opt-ins |
Stage 0 — state detection sweep
Read all of the on-disk signals at once. Each subsequent stage will skip its question(s) when the corresponding signal is already present.
ls .game-of-cards/deck/ 2>/dev/null && echo "DECK_EXISTS" || echo "deck_missing"
which goc 2>/dev/null && echo "GOC_ON_PATH" || echo "goc_missing"
grep -l '<!-- BEGIN GOC' AGENTS.md CLAUDE.md CLAUDE.local.md 2>/dev/null && echo "BRIEFING_MERGED" || true
grep -E '^autonomy:' .game-of-cards/config.yaml 2>/dev/null && echo "AUTONOMY_SET" || true
[ -d .git ] && echo "GIT_REPO" || echo "git_missing"Route on the results:
DECK_EXISTS+AUTONOMY_SET→ report "GoC is already initialized in this repo — deck is live." and stop. Do not continue. This is the silent-exit branch the skill description promises.DECK_EXISTSwithoutAUTONOMY_SET→ install completed previously but the autonomy question was deferred. Skip Stages 1–5 and jump directly to Stage 6 so the user can pick a mode now.- The sweep, or a later
goccall, is denied by the host's permission check ("Permission to use Bash has been denied", "requires approval") →gocmay be installed, just not allowed; do not report it missing. Invoke the host's kickoff complement for the grant —claude-kickoffaddsBash(goc:*)topermissions.allowin its Stage 3 — then re-run kickoff. goc_missing→ halt at this stage. Tell the user: "gocCLI is not installed. Install it (e.g.pipx install game-of-cards, or use whatever plugin/runtime ships goc for your host) and re-run kickoff." Do not proceed withoutgoc.git_missing→ surface this one-line notice once, then continue: "No git repository here — version control is not set up. The deck assumes git (auto_commit, claim history, closure logs); rungit initif you want the deck tracked." Do not explain further; the user decides whether togit initbefore or after kickoff completes.- Otherwise → continue. Hold the detected flags in mind through the rest of the flow.
Stage 1 — introduce GoC
Skip this stage if BRIEFING_MERGED was detected in Stage 0 — it means
the user has engaged with kickoff before, and replaying the intro is
noise.
Otherwise, deliver this paragraph verbatim (no edits, no summarising):
Game of Cards (GoC) is a lightweight, file-based methodology layer for software projects. It keeps a deck of cards — each card is a plain Markdown file with YAML frontmatter — that records every piece of persistent work: features, bugs, decisions, and experiments. The deck tracks status, value, and dependency edges between cards, so AI agents and human contributors share one authoritative view of what's in flight and what's next. GoC borrows from XP (user stories, spike experiments), Kanban (pull-based intake, explicit WIP limits), and Scrum (definition of done, backlog refinement). The CLI (
goc) validates cards, renders board views, and gates closure on a per-card definition of done.
Then pause — do not continue to Stage 2 until the user acknowledges or responds.
Stage 2 — persona question
Skip this stage if BRIEFING_MERGED was already detected (Stage 3 has
nothing left to ask, so persona is moot).
Otherwise, ask the user one question:
Which description best fits how you'll use GoC here?
- Solo / vibe-coder — I'm the only contributor; I want a lightweight personal task tracker with AI-agent help.
- Classical team — We have a shared repo, multiple humans, and want the deck as a shared source of truth alongside PRs and reviews.
- OSS / library evaluator — I'm exploring GoC before deciding whether to adopt it; I'd prefer to keep the repo footprint minimal.
- Agent-runtime / CI — This repo is driven primarily by AI agents; I want the CLI plus a runtime briefing that any agent can read.
Record the user's answer. It drives the briefing-target recommendation in Stage 3.
| Persona | Recommended briefing target |
| Solo / vibe-coder | CLAUDE.local.md — gitignored, no commit footprint |
| Classical team | CLAUDE.md — checked-in Claude Code guidance (or AGENTS.md if cross-runtime visibility matters) |
| OSS / library evaluator | AGENTS.md — cross-runtime visible, minimal Claude assumption |
| Agent-runtime / CI | AGENTS.md — read by every modern agent runtime |
Stage 3 — pick a briefing home
Skip this question if Stage 0 detected BRIEFING_MERGED — read the
existing target off disk (grep -l '<!-- BEGIN GOC' AGENTS.md CLAUDE.md
CLAUDE.local.md) and pass that file to Stage 4.
Otherwise, ask the user one question, ordering the options by persona but always offering all three:
Where should the GoC briefing live in this repo? The briefing is one marker-bounded
<!-- BEGIN GOC -->block — agent-neutral runtime instructions plus, when CLAUDE.md is the home, the Claude-specific setup notes inline. Futuregoc upgradere-syncs only the chosen file.
- AGENTS.md — read by Codex, Cursor, Copilot, OpenCode, Aider, and Claude Code (via the
@AGENTS.mdimport thatgoc install/goc upgradewrites into CLAUDE.md when Claude is installed). Recommended for cross-runtime, agent-runtime, and OSS-eval personas.- CLAUDE.md — read only by Claude Code; the host-agnostic body plus the Claude-specific extras live inline. Cross-runtime visibility is given up (Codex/Cursor/etc. won't see the briefing). Recommended for Claude-only teams.
- CLAUDE.local.md — read only by Claude Code, gitignored by default. Recommended for solo/personal use where the deck is private and you don't want a checked-in briefing.
Lead with the persona's recommended option but show all three. Record
the answer (one of AGENTS.md, CLAUDE.md, CLAUDE.local.md) for
Stage 4. If the persona is Classical team, follow with the tip in
reference.md § Team tip: external deck location.
Stage 4 — scaffold project state
Summarise the briefing-target answer collected in Stage 3, then ask:
Ready to scaffold? This will create
.game-of-cards/and merge the GoC briefing into<chosen file>. Confirm? [yes/no]
On confirmation, run goc install --briefing-target <chosen file> (the
flag tells the install primitive which file holds the briefing block;
omit it only when the user wants the default AGENTS.md).
Note for this repo (game-of-cards source tree): Translate bare
goccommands touv run gocwhen working in the goc package source tree itself.
The command writes project state and merges the briefing block into
the chosen file (creating it if absent) — full inventory and the
CLAUDE.md-import behaviour in reference.md § What gets installed.
After goc install returns, verify .game-of-cards/deck/ exists and
the chosen file contains a <!-- BEGIN GOC --> block before
continuing.
Stage 5 — confirm ready
Report to the user:
GoC is set up. The deck is live; `goc` and `goc validate` work.Continue to Stage 6.
Stage 6 — pick an autonomy mode
Skip this stage if Stage 0 detected AUTONOMY_SET — the user already
chose a mode on a previous run and re-asking is noise.
Otherwise, ask the user one question:
How would you like cards to be pulled off the queue? GoC's whole point over a plain task list is that the deck is designed to be drained autonomously when
human_gate: none. Pick a mode (you can change later by editing.game-of-cards/config.yaml):
- Manual — you file cards and work them by hand. No setup.
- Supervised loop — your host runs
pull-cardon a timer inside a session you watch (e.g./loop 30m /pull-cardon Claude Code). Best for trying GoC before automating further.- Local cron — a cron job runs
pull-cardunattended on your machine. Best for solo workflows that trust the agent to drainhuman_gate: nonecards overnight.- CI / GitHub Action — a workflow runs
pull-cardon a schedule in CI. Best for shared decks across a team.- Skip for now — defer the choice; kickoff will ask again on next run.
Record the answer in .game-of-cards/config.yaml as a top-level
autonomy: key with one of these values: manual, loop, cron,
action. For the Skip for now option, do not write the key — its
absence is what makes the next kickoff re-ask. Stage 0 detects the
written key on re-run and skips this stage.
The key is a record, not a setup: no goc command reads it (only Stage
0's re-run check does), so loop, cron or action drains nothing
until a scheduler is wired. If the host has its own kickoff complement
(Claude Code ships claude-kickoff, Codex codex-kickoff, OpenClaw
openclaw-kickoff), invite the user to run it now — its autonomy
recipes wire each mode on that host, and say so where the host has
none. Otherwise recommend the next step:
What should the first card be?The deck is now live. All other GoC skills work immediately — no further generic kickoff needed.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

