Chalk Labs
Need help deploying co-thinker?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/bodnarbalazs-co-thinker
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: testing, productivity
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md and code. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/co-thinker
What's inside
- Skill name
co-thinker- Description
- An idea not ripe yet. Use when a thought, a doubt or a preference is still open, when a question surfaces that is not yet a ticket, or when a trade-off needs testing before intent hardens into a plan or a Decision Record.
- License
- MIT
- Instructions
- 53 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# Co-Thinker Think alongside the user until the choices in front of him are visible and the thinking has a home. ## Must-Reads 1. From the repository root, read `dydo/understand/about.md`. 2. From the repository root, read `dydo/understand/architecture.md`. 3. From the repository root, read `dydo/glossary.md`. 4. From the repository root, read `dydo/reference/linear-workspace-standard.md`. ## Boundary Think is the stage: a raw idea arrives from the user and leaves ripe, leaves as a Decision Record, or waits; the route is the map holder's to draw. Every fact the repository, the environment or a primary source can supply is yours to fetch before you ask the user anything; what reaches him is what he alone has the authority to settle. ## Method 1. **Do the homework first.** Read the code, the Decision Records, the plan and the glossary that already cover this ground; send `research` after anything outside the repository, naming the ticket as its destination once one exists, and take the one-line answer it reports back. Done when every question still standing is one only the user can answer. 2. **Name the real choice.** Separate what is known from what is assumed, then frame the live options with what each one buys and what it costs, drawn by `show-me` when a shape says it faster than prose. Done when every option carries its trade-off and nothing assumed is dressed as settled. 3. **Put the open choices to the user.** Use `grilling`: a round at a time, each question carrying your recommendation; a question prose cannot settle goes to `prototype`. Done when no branch of the idea is left unvisited and nothing rests on a silent assumption. 4. **Fix the words.** When the exchange keeps sliding on a term, run `domain-modeling`: one name per concept, written into `dydo/glossary.md` where the next session finds it. Done when both sides use one word for one thing. 5. **Test the edges.** Try the idea against a concrete example and a counterexample, and ask what can be removed. Done when it survives both, or has changed shape to survive them. 6. **Recommend, then close.** Give a reasoned preference instead of handing the choice back as a list, and say what is settled, what is still open, and where each piece lands. Done when the user can see all three. ## Handoff | What leaves | Where it lands | |---|---| | Hard to reverse, surprising later, a real trade-off | a Decision Record in `dydo/project/decisions/` | | An unscheduled strategic possibility worth preserving | a Linear ticket in `FutureFeature` | | Shared feature intent within a Project | the map holder maintains its anchor following `dydo/reference/feature-spec-convention.md`, read from the repository root | | Project intent after the homework | the user invokes `to-project`, creating a Project in `Backlog` | | Ripe atomic ticket intent | file the ticket in `Todo` with one Type, one Mode, outcome, owned paths, blockers, exact gates and base branch; the user opens a captain session or spawns one from their session; this session does not become the captain | A FutureFeature stays one until the user promotes it. Hand over what is written down, the Linear ticket or Project, the Decision Record, the glossary entry, never this conversation.
From bodnarbalazs/dydo, MIT.
More from bodnarbalazs/dydo
- 0.622score
overnight-mode
bodnarbalazs
Overnight mode, the user away while admirals run on standing and night orders.
SkillAFreshness: A★ 6Claude Code +9 - 0.622score
maintain-formal-verification-skill
bodnarbalazs
Bring a project's formal verification back in line with the code after the code changes.
SkillAFreshness: A★ 6Claude Code +9 - 0.622score
maintain-verification-skill
bodnarbalazs
Keep a project's verification skill and feature map true to the running app.
SkillAFreshness: A★ 6Claude Code +9 - 0.622score
retro
bodnarbalazs
Kaizen for the harness.
SkillAFreshness: A★ 6Claude Code +9 - 0.619score
wizard
bodnarbalazs
Human-only setup, credentials or cutover steps need a repeatable interactive bash procedure; the Enablement captain uses this method.
SkillAFreshness: A★ 6Claude Code +9 - 0.617score
in-your-own-words
bodnarbalazs
Restate in your own words what you think the user's goals are and what problem they're trying to solve.
SkillAFreshness: A★ 6Claude Code +9 - 0.608score
bro
bodnarbalazs
Stop.
SkillAFreshness: A★ 6Claude Code +9 - 0.597score
domain-modeling
bodnarbalazs
Build and sharpen a project's domain model.
SkillAFreshness: A★ 6Claude Code +9 - 0.597score
diagnosing-bugs
bodnarbalazs
Diagnosis loop for hard bugs and performance regressions.
SkillAFreshness: A★ 6Claude Code +9 - 0.597score
ketchup
bodnarbalazs
Catch the user up on what changed in this chat since their last message, in a short reply built on real visuals.
SkillAFreshness: A★ 6Claude Code +9 - 0.597score
prototype
bodnarbalazs
Build a throwaway prototype to answer a design question.
SkillAFreshness: A★ 6Claude Code +9 - 0.597score
grilling
bodnarbalazs
Grill the user relentlessly about a plan, decision, or idea.
SkillAFreshness: A★ 6Claude Code +9
Signals
- Stars
- 6
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Oct 24, 2026 (11 days ago)
- Last release
- Oct 19, 2026 (15 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/bodnarbalazs/dydo
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from bodnarbalazs
- 0.622score
retro
bodnarbalazs
Kaizen for the harness.
SkillAFreshness: A★ 6Claude Code +9 - 0.622score
maintain-formal-verification-skill
bodnarbalazs
Bring a project's formal verification back in line with the code after the code changes.
SkillAFreshness: A★ 6Claude Code +9 - 0.622score
maintain-verification-skill
bodnarbalazs
Keep a project's verification skill and feature map true to the running app.
SkillAFreshness: A★ 6Claude Code +9 - 0.622score
overnight-mode
bodnarbalazs
Overnight mode, the user away while admirals run on standing and night orders.
SkillAFreshness: A★ 6Claude Code +9 - 0.619score
wizard
bodnarbalazs
Human-only setup, credentials or cutover steps need a repeatable interactive bash procedure; the Enablement captain uses this method.
SkillAFreshness: A★ 6Claude Code +9 - 0.617score
in-your-own-words
bodnarbalazs
Restate in your own words what you think the user's goals are and what problem they're trying to solve.
SkillAFreshness: A★ 6Claude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get co-thinker running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
co-thinker skill. Agent Mart. https://agentmart.dev/skills/bodnarbalazs-co-thinker (last updated 2026-10-24)