michi-loop SKILL.md skillA

michi-loop from csepulv/michi is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

Install it, pinned
mdr add csepulv/michi/michi-loop@git:20261106.9667115

Pins git:20261106.9667115 in mdr.lock. Pin the fingerprint instead.

Audit grade
A, 17 of 17 checks passed
Kind
SKILL.md skill
Repository
csepulv/michi
Path
toolkit/skills/michi-loop/SKILL.md
Versions recorded
4
Latest version
git:20261106.9667115, committed Nov 6, 2026
First version recorded
Jul 13, 2026
Size
11,596 bytes
Repository stars
2
License
MIT

Set up and launch a disciplined lights-off loop — gate readiness, define the goal/verifier/stop contract, hand off to a loop mechanism, review at halt. For convergent work with a verifier that can be trusted. The convergent + autonomous counterpart to michi-expedition.

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

# Michi Loop

Run a task as an autonomous, lights-off **loop** — the agent works unattended, checks its own output
against a verifier, and stops on a defined condition — *with Michi discipline around it.*

This skill is **discipline over mechanism**. It does not run the loop; Claude Code already ships the
runners (`/loop`, `/goal`, `claude -p`, `Workflow`). What it adds is the judgment the loop-engineering
literature skips: *have you earned lights-off, what does "done" mean, what checks the work, and when
does it stop?* You are invoked once, as the **designer and launcher** of a loop — not its body.

It is the **convergent + autonomous** lane — the sibling of `michi-expedition` (emergent +
Entrusted-spiral). The boundary between them is the *form of verification you can answer* (see the Gate).

**Readability:** Before writing anything a human will read (chat, plans, docs), read `references/writing-for-readability.md`.

**Principles:** Before starting, read `references/principles.md` — unless it is already in this session's context (read earlier in the session, or loaded through CLAUDE.md).

## When to use (and when not)
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add csepulv/michi/michi-loop@sha256:861d61f640a768d2

The line at the top of this page pins the label git:20261106.9667115, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_7zy4aaifwyxw6vjo.svg)](https://markdownregistry.com/a/art_7zy4aaifwyxw6vjo)

2 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 4 versions of this file, between Jul 13, 2026 and Nov 6, 2026, about one every 38 days.

versioncommittedcommitsizeaudit
git:20261106.9667115 latest2026-11-06 9667115 11,596 BA view · diff
git:20261029.22b5d332026-10-29 22b5d33 11,397 BA view · diff
git:20260921.b6353772026-09-21 b635377 11,270 BA view · diff
git:20260713.1de48182026-07-13 1de4818 10,932 BA view

Audit of the latest version

A  17 of 17 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (11596 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

csepulv/michi · 2 stars · license MIT · pushed 2026-11-06 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_7zy4aaifwyxw6vjo
GET https://markdownregistry.com/api/v1/resolve?ref=csepulv/michi/michi-loop
GET https://markdownregistry.com/api/v1/blob/861d61f640a768d24146d6eb9d4c1908a8a553b4cc61714a0396bfc071966a15

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.

More from csepulv/michi

michi-away SKILL.md skill
csepulv/michi · toolkit/skills/michi-away/SKILL.md · Switch to Entrusted while the human steps away — keep running the Michi cycle (debrief, planning, session) on a…
git:20261106.9667115 · audit A · 2 stars
michi-bootstrap SKILL.md skill
csepulv/michi · toolkit/skills/michi-bootstrap/SKILL.md · Onboard a project to the Michi documentation structure — survey what exists, recommend what to build (S/M/L/XL effort…
git:20261106.9667115 · audit A · 2 stars
michi-debrief SKILL.md skill
csepulv/michi · toolkit/skills/michi-debrief/SKILL.md · Structured post-session review — assess what was delivered, review decisions, capture learnings, audit what got…
git:20261106.9667115 · audit A · 2 stars
michi-diagrams SKILL.md skill
csepulv/michi · toolkit/skills/michi-diagrams/SKILL.md · Draw Mermaid diagrams for project documentation, design docs, READMEs, specs and architecture pages that will be viewed…
git:20261029.22b5d33 · audit A · 2 stars
michi-docs-site SKILL.md skill
csepulv/michi · toolkit/skills/michi-docs-site/SKILL.md · Scaffold an internal Astro + Starlight docs browser that renders a project's markdown as a searchable site, or write…
git:20261106.9667115 · audit A · 2 stars
michi-expedition SKILL.md skill
csepulv/michi · toolkit/skills/michi-expedition/SKILL.md · Run a Michi expedition — open-ended, Entrusted, spiral exploration where the end isn't known (data exploration…
git:20261106.9667115 · audit A · 2 stars
michi-explore SKILL.md skill
csepulv/michi · toolkit/skills/michi-explore/SKILL.md · Structured conversation with Michi discipline — explore ideas, research options, orient in a new space. Follows the…
git:20261106.9667115 · audit A · 2 stars
michi-library SKILL.md skill
csepulv/michi · toolkit/skills/michi-library/SKILL.md
git:20261106.9667115 · audit B · 2 stars
michi-planning SKILL.md skill
csepulv/michi · toolkit/skills/michi-planning/SKILL.md · Prepare a milestone for a Michi implementation session — write the plan doc, define acceptance criteria, co-design…
git:20261106.9667115 · audit A · 2 stars
michi-pr-prep SKILL.md skill
csepulv/michi · toolkit/skills/michi-pr-prep/SKILL.md · Prepare a PR review guide — a companion document that helps reviewers understand what changed, why decisions were made…
git:20261106.9667115 · audit A · 2 stars
michi-scenario-test-builder SKILL.md skill
csepulv/michi · toolkit/skills/michi-scenario-test-builder/SKILL.md
git:20261106.9667115 · audit B · 2 stars
michi-second-look SKILL.md skill
csepulv/michi · toolkit/skills/michi-second-look/SKILL.md · Independent review of a milestone, epic or project by an agent that was not in the session — a second pair of eyes…
git:20261106.9667115 · audit A · 2 stars

Every file in csepulv/michi

Browse by kind, by grade A, or by owner.