harness-loop
Skill3 starsChanged 27 days agoMITView on GitHub
Run a large goal as a loop with one check, a progress file on disk and a turn cap. Uses the deliver workflow for multi-file work. Use when the user asks you to keep going until a big goal is finished.
5 other projects use a version of this.
Adopted by
Most recently taken up by
- niksavis/handilytheir own version, first seen Nov 8, 2026
- niksavis/basiclytheir own version, first seen Nov 7, 2026
- anyoneanderson/agent-skillstheir own version, first seen Oct 13, 2026
- Chachamaru127/claude-code-harnesstheir own version, first seen Oct 13, 2026
- z3e8/pyharnesstheir own version, first seen Oct 11, 2026
Other versions5 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 5
- Chachamaru127/claude-code-harness★ 3.2ksame name
- anyoneanderson/agent-skills★ 3same name
- niksavis/basicly★ 2same name
- niksavis/handily★ 2same name
- z3e8/pyharness★ 0same 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 repository25 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- adrSkill
- api-contract-syncSkill
- architecture-advisorSkill
- ci-triageSkill
- compassSkill
- debugSkill
- delivery-digestSkill
- design-taste-skillsSkill
- grillSkill
- harness-improveSkill
- impact-analysisSkill
- native-mobile-skillsSkill
- plugin-authorSkill
- pr-flowSkill
- project-memorySkill
- remaining-workSkill
- reviewSkill
- slide-deckSkill
- spec-flowSkill
- task-flowSkill
- ui-conformanceSkill
- working-style-skillsSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
19 rules. 0 are common, 17 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
17
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 17
Check: one command that proves the goal.
- Distinctive: line 18
Stop condition: the check passes, the turn cap is reached, or a blocker stops the work.
- Very specific: line 22
Write the progress file .agents/progress/<goal-slug>.md.
- Distinctive: line 23
Read it at the start of each turn.
- Distinctive: line 27
Read the repository rules, the project memory and the specs that the goal touches.
- Distinctive: line 31
Do one phase in each turn.
- Distinctive: line 32
After each phase, read your diff and run the check.
- Distinctive: line 34
Write the result in the progress file.
- Distinctive: line 39
Run 2 workers at most. Give them files that do not overlap.
- Distinctive: line 40
Before the first edit in a worktree or a new branch, run git log --oneline main..HEAD and git log --oneline HEAD..main.
- Distinctive: line 41
If the branch is behind and has no commits of its own, run git merge --ff-only main.
- Distinctive: line 42
If the branch is behind and has its own commits, tell the parent.
- Distinctive: line 43
If the user allows commits, commit each finished piece.
- Distinctive: line 45
Do not remove a worktree because it looks finished.
- Distinctive: line 51
Run the shared steps in the session worktree: merge, screenshots, land.
- Distinctive: line 53
Read a saved subagent report for "the hook refused" notes before you trust it.
- Distinctive: line 57
Do not commit, push, open a PR, deploy, publish, write to production or run a migration on production without approval.
- Very specific: line 58
Use only commands that exist: in package.json, in a Makefile, or in the documented command list.
- Distinctive: line 59
Prefer a narrow fix to a broad rewrite.
“Common” is measured: kept by at least 6 in 10 of the 4 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
harness-loop
Set up
State these items before you start. Ask only about an item that blocks safe work.
- Goal: one sentence.
- Check: one command that proves the goal.
- Stop condition: the check passes, the turn cap is reached, or a blocker stops the work.
- Turn cap: 20 turns unless the user says another number. Never raise the cap yourself.
- Out of scope: what you will not touch.
Write the progress file .agents/progress/<goal-slug>.md. It holds the five items above, the done list, the next step and the blockers.
Read it at the start of each turn. A new context resumes from this file.
Loop
- Read the repository rules, the project memory and the specs that the goal touches.
- For work on more than one file, run the
deliverworkflow (.claude/workflows/deliver.js) with the goal as args. The workflow plans, then splits the work into 2 file groups at most. Workers build. A reviewer checks in a fresh context. Fixes take 2 rounds at most. Without workflow support, do the same steps with thescout,workerandrevieweragents. - Do one phase in each turn.
- After each phase, read your diff and run the check.
- Fix a failure that you introduced before you start the next phase.
- Write the result in the progress file.
- At the cap, report "partial" with the blocker. Do not continue.
Parallel work
- Run 2 workers at most. Give them files that do not overlap.
- Before the first edit in a worktree or a new branch, run
git log --oneline main..HEADandgit log --oneline HEAD..main. - If the branch is behind and has no commits of its own, run
git merge --ff-only main. - If the branch is behind and has its own commits, tell the parent. Do not rebase.
- If the user allows commits, commit each finished piece. A session that dies loses uncommitted work.
- Stay inside the files that your task names.
- Do not remove a worktree because it looks finished. Wait for the signal of the orchestrator.
Worktree fence
The desktop app fences each subagent. A hook refuses Edit, Write and git commands that target another worktree.
- Run the shared steps in the session worktree: merge, screenshots, land.
- A reviewer of a lane worktree calls
EnterWorktreewith the lane path first. - Read a saved subagent report for "the hook refused" notes before you trust it. A refusal appears in the notes, not as a failure.
Safety
- Do not commit, push, open a PR, deploy, publish, write to production or run a migration on production without approval.
- Use only commands that exist: in
package.json, in a Makefile, or in the documented command list. Do not invent one. - Prefer a narrow fix to a broad rewrite.
Handoff
Give the goal, the files changed, each command with its result, the risks and the next step.
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.

