north-star
Skill0 starsChanged 36 days agoMITView on GitHub
Print one screen for the current project, with the vision, what's built and open against each goal, and the drift.
14 other projects use a version of this.
Adopted by
Most recently taken up by
- Glenf/skillstheir own version, first seen Nov 15, 2026
- uzysjung/uzys-agent-harnesstheir own version, first seen Nov 8, 2026
- lifekit-hq/devclawtheir own version, first seen Oct 21, 2026
- andreaskelm/pm-braintheir own version, first seen Oct 17, 2026
- fusebase-dev/fusebase-flowtheir own version, first seen Oct 16, 2026
- agentmart/pm-ostheir own version, first seen Oct 15, 2026
Used with
Projects with it also use strategy.
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.
EpistatesSkillstrategy
Plan weekly or monthly X content strategy. Use when the user wants to plan their content calendar, set growth goals, build a content pipeline, or think strategically about their X presence.
Also in 3 projects that use this
Other versions14 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 14
- andreaskelm/pm-brain★ 235same name
- fusebase-dev/fusebase-flow★ 9same name
- ai-analyst-lab/north-star★ 8same name
- BhaskarManam/pstack★ 4same name
- raphaelDkhn/worldbuilder★ 3same name
- uzysjung/uzys-agent-harness★ 3same name
- ele-yufo/unstuck★ 2same name
- lifekit-hq/devclaw★ 2same name
- Naoray/skills★ 2same name
- poponline63/north-star★ 2same name
4 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 repositoryOne other file it gives its agents.
- specifySkill
The file
We split this skill into the rules it gives an agent
5 rules. 0 are common, 4 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 12 edited versions we read keep these.
4
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
Every line is a claim you checked against its source.
- Distinctive: line 33
Run git fetch --quiet first so the default branch matches the remote.
- Distinctive: line 51
done: when its scope is built or its signal is met.
- Very specific: line 52
moving: when it has an open PR or a ticket someone has committed to (todo, in progress, in review).
- Distinctive: line 53
idle: when planned scope remains and nothing is moving it.
“Common” is measured: kept by at least 6 in 10 of the 12 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
North star
Print one screen that answers "where are we against the vision?" for the project in the working directory. A reader who lost the thread five sessions ago finishes it knowing the goals, what serves each one, and what has wandered off.
Every line is a claim you checked against its source. A claim you can't check stays off the screen.
1. Find the vision doc
The vision doc says who the product is for, what they get, and how success is judged, ideally with dates. Take the first of these that exists:
- The docs
CLAUDE.mdorAGENTS.mdpoint to for the vision, the bet, the goals, the scope or the milestones. Several docs can split the job, one holding the bet and another the milestones. Read every one of them. - A file named like
NORTH_STAR.md,VISION.md,BET.md,PRODUCT.mdorROADMAP.md, at the root or underdocs/.
A README that only describes what the code does is not a vision doc. With neither, run INTERVIEW.md, then continue here.
Done when you can name the vision doc paths and quote the sentence that states the north star.
2. List the goals
Goals are the dated or measurable outcomes the vision doc names: milestones, success signals, secondary targets. Use the doc's own ids (M1), or G1, G2 when it has none. A success signal that a milestone already proves folds into that milestone. For each goal, record its id, a short name, its due date and the scope it lists.
Done when each goal traces to a line in the doc.
3. Gather the work
Run git fetch --quiet first so the default branch matches the remote.
- Built.
git log --first-parent --format='%h %as %s' -n 300on the default branch. Squash and merge titles carry the PR number. Past 300 commits, say so in the sources line. The commits that wrote the vision docs are the vision, not built work. - Open. Open PRs from
gh pr list --state open. Open tracker issues: find the tracker from issue ids in commit titles and branch names (ENG-65,#123) and from tracker names inCLAUDE.md,AGENTS.mdand the docs. Query it through its MCP server or CLI, scoped to this project. Every state other than done or canceled counts as open. - Planned. Scope a goal lists that has no built work and no open PR or ticket. Unchecked
- [ ]items in the vision doc and the docs it links are planned too, and go on the Checklist row. A checkbox is a plan, not motion.
A source you can't reach (no gh auth, no tracker MCP) is a gap. Name it in the sources line.
Docs go stale faster than git and the tracker. When a doc's status line or checklist disagrees with the record (it says "not built" and a merged PR built it), the record wins. Name the stale doc in the sources line.
Done when every commit in the range and every open item has been read.
4. Tie each item to a goal
Tie work to the goal whose text it advances: a ticket that names the milestone, a PR that builds what the goal lists. Enabling work (contracts, fences, CI, docs) ties to the goal it unblocks. Work that ties to no goal's text is drift. A tie that needs stretching is drift too.
Give each goal one status:
- done when its scope is built or its signal is met.
- moving when it has an open PR or a ticket someone has committed to (todo, in progress, in review). A backlog ticket is open but parked. Mark it
(backlog)in the Open row. - idle when planned scope remains and nothing is moving it. An idle goal is drift.
Done when every item has exactly one goal or sits in drift.
5. Print the screen
Print markdown, outside any code block, so the terminal renders the headings, bold and lists. Keep it to one screen: one heading per goal and at most four one-sentence bullets under it.
# <Project>
<The north star in one or two sentences.>
**Now:** <What is in flight, the nearest goal, and what stands between it and done. Two sentences at most.>
### <id> · <name> · <Mon D> · <n> days · <done | moving | idle>
- **Built:** <features, named the way a user would name them>
- **Open:** <open PRs and tickets, each with its id>
- **Left:** <scope with nothing built or open>
- **Checklist:** <count of unchecked items, then the soonest and its date> (<file>)
### Drift
- <idle goals, soonest due first>
- <work that serves no goal>
*Sources: <docs read> · <n> commits since <date> · <tracker>, <open> of <total> open · <gaps and stale docs>*- Order goals by due date and count days from today.
- Ids appear only on open items (
#43,ENG-66), where the reader acts on them. The commits behind a built feature stay in your check. - Drop a bullet that has nothing to report, rather than writing one that says so.
Before printing, check every claim against its source. Each built feature traces to the commits that built it, each open id exists in the state you claim, each date and goal name matches the doc, and each tie traces to the goal's text. Fix or cut what fails, then print the screen and nothing else.
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.

