harness-onboarding
Skill2 starsChanged 5 months agoMITView on GitHub
Bootstrap pack for a new developer joining the project — reads context-packs, specs, recent decision-log entries, and decisions-locked, then produces a role-tailored onboarding pack (executive summary + sequenced reading + first-task suggestions). Reduces time-to-first-PR.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- Makkov-design/makkov-harnesstheir own version, first seen Nov 15, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- Makkov-design/makkov-harness★ 8same name
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 repository24 other files it gives its agents.
- AGENTS.mdAGENTS.md
- control-plane/AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- control-plane/CLAUDE.mdCLAUDE.md
- agentic-os-reference-syncSkill
- branch-cleanupSkill
- capture-triageSkill
- darwin-housekeepingSkill
- darwin-path-modeSkill
- decision-log-entrySkill
- eodSkill
- expert-interview-guideSkill
- feedback-intakeSkill
- ingest-contentSkill
- meeting-to-work-itemsSkill
- memory-consolidateSkill
- morning-briefSkill
- os-bootstrap-extendSkill
- os-bootstrapSkill
- pr-reviewSkill
- scheduleSkill
- spec-cross-checkSkill
- ux-critiqueSkill
- weekly-retroSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 4 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
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 19
If the project has neither a harness nor a README, the skill asks the principal whether to construct an ad-hoc pack from git log + …
- Distinctive: line 57
Do not dump everything. For each role, prioritize:
- Distinctive: line 75
Other role: ask the principal to map → do not improvise.
- Very specific: line 79
Read decisions-locked.md and look for entries with status pending or known blockers.
- Very specific: line 86
Read docs, open a PR adding OPEN_QUESTIONS.md to src/<module>/ capturing initial questions.
- Distinctive: line 91
Read decision-log + decisions-locked → entry "post-reading review" in the decision-log (verifies comprehension).
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Skill: harness-onboarding
Preconditions
- The project has either a
docs/harness/structure (preferred) or, at minimum, aREADME.md+ARCHITECTURE.md+ recent commits the skill can read. - The principal has supplied
dev_nameandrole. - A target location for the output pack is writable (defaults to
docs/harness/onboarding-<dev_name>-YYYY-MM-DD.md).
If the project has neither a harness nor a README, the skill asks the principal whether to construct an ad-hoc pack from git log + top-level files, or to defer until docs exist.
When to use
A new developer is joining the project and needs to absorb context before contributing. Without a structured pack, week 1 is spent asking the obvious instead of producing a PR.
NOT for:
- Re-briefing an existing dev on a new scope → use
pr-reviewor direct conversation. - Onboarding a non-technical stakeholder → produce a different artifact (slide deck, not technical pack).
Inputs
dev_name(required) — the developer's name.role(required) — expected role (owner of module X, contributor to layer Y, etc.).project(optional) — the project subfolder; defaults to repo root.depth(optional) —"essential"(~1h reading) or"deep"(~1 day reading).
Sequence
Step 1 — Map harness inputs
For a project with a docs/harness/ structure, locate and read:
docs/harness/README.md— entry pointdocs/harness/context-packs/project-state.md— current statedocs/harness/context-packs/decisions-locked.md— locked premisesdocs/harness/context-packs/stakeholders.md— who is whodocs/harness/context-packs/<domain>.md— business vocabularydocs/harness/decision-log.md— recent decisions (last 20)docs/harness/specs/<relevant-to-role>.md— spec of the module / feature the dev will own
For projects without a formal harness, read the equivalents: README, ARCHITECTURE.md, recent commits, OPEN_QUESTIONS.md, any decisions/ folder.
Step 2 — Filter by role
Do not dump everything. For each role, prioritize:
Module owner:
- The canonical spec for the module the dev will own.
- Operational docs in
src/<module>/docs/*. - Decisions-locked entries that fix the design.
- Adjacent specs that the module consumes or produces.
Pipeline / DS owner:
- Data layer READMEs (raw → preprocessed → domain → features).
- Decisions-locked about data conventions.
- Spec-free convention docs (pipeline is convention, not module).
Harness / governance owner:
- Full
docs/harness/README.md. - All
workflows/andsignals/docs. - The entire decision-log (to learn the entry pattern).
Other role: ask the principal to map → do not improvise.
Step 3 — Identify current blockers
Read decisions-locked.md and look for entries with status pending or known blockers. The output: a list of what the new dev does not need to resolve (already in flight) and what is waiting externally.
Step 4 — Map first tasks
By role, suggest 2-3 realistic first PRs:
Module owner:
- Read docs, open a PR adding
OPEN_QUESTIONS.mdtosrc/<module>/capturing initial questions. - Set up local environment and run the existing pipeline end-to-end with a dummy snapshot.
- First skeleton of
build()calling the correct inputs (even without real logic).
Harness / governance owner:
- Read decision-log + decisions-locked → entry "post-reading review" in the decision-log (verifies comprehension).
- Audit context-packs vs decision-log for drift.
- Propose one harness improvement via PR.
Step 5 — Output: onboarding pack
# Onboarding pack — <dev_name>
**Role:** <role>
**Project:** <project>
**Estimated reading:** ~Xh (depth=<>)
## Sequenced reading
1. <file> — <why this first>
2. <file> — <why next>
...
## Locked premises that affect <dev_name>
- D-NNN: <decision> — relevance to role
- D-NNN: ...
## Known blockers (not <dev_name>'s problem)
- B-NN: <blocker> — owner / waiting on
## First three PRs
1. <PR scope> — <expected outcome>
2. ...
3. ...
## Who to ask what
- <person> — <area of authority>
## Anti-patterns to avoid (project-specific)
- <pattern> — why it bitesStep 6 — Hand off
Save as docs/harness/onboarding-<dev_name>-YYYY-MM-DD.md and share the link. Offer a 30-min walkthrough as a follow-up if the dev wants it.
Anti-patterns
- Dumping the whole repo — defeats the purpose. Filter aggressively by role.
- Skipping the blockers section — new devs accidentally re-solve already-resolved problems if they don't know the lock-in.
- Vague first tasks — "explore the codebase" is not a first task. Give 2-3 concrete PRs the dev can ship in week 1.
- No "who to ask" section — the dev hits a question, has no idea who owns the answer, posts in the wrong channel.
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.

