Skip to content
Get packs

Guide · Agent skill

How to use the Unify Harmony Checker agent skill

Updated · written with AI from the project's own documentation and checked against our catalogue data

Unify is an agent skill that checks whether a set of things you care about, such as systems, rules, skills, goals, documents or plans, pull in the same direction. It is for people who run multi-part personal or team systems, or who build agent setups where the parts need to stay consistent. By the end of this guide you will know what the skill does, how to install it, how to ask for a check in plain language, and what it needs to run.

Unify Harmony Checkerdanielmiessler/LifeOS · MIT

Quick start

  1. 1Download and unzip. Download the Unify ZIP from this page and unzip it.
  2. 2Place the folder. Move the Unify folder into ~/.claude/skills/ or a project's .claude/skills/ folder, or upload the ZIP under Settings, Capabilities, Skills in the Claude apps.
  3. 3Install Bun. Make sure the Bun runtime is installed, since the CheckReport.ts script runs under bun.
  4. 4Set up the default config if needed. For a run with no inputs, create Config.md in USER/CUSTOMIZATIONS/SKILLS/Unify/ inside your LifeOS install.
  5. 5Ask in plain language. Use a trigger phrase such as are these in harmony, naming the inputs you want checked.
  6. 6Review the ranked recommendations. Read the conflicts and recommendations, decide which are real, and approve any changes yourself.

What the Unify skill makes the agent do

Unify takes a set of inputs and works out what each one is for. It extracts each input's core principles and its ultimate aim, then maps how the inputs relate. Some reinforce each other, some pull in productive tension, some duplicate each other, and some work against each other.

It runs the FirstPrinciples, Superior and SystemsThinking skills, along with other thinking skills where they fit, to look for real conflicts. It then produces a short ranked list of recommended changes meant to bring the whole into unity. The skill is read-only: it recommends and never edits your files.

  • Extracts each input's principles and aim
  • Maps reinforcing, tense, drifting, duplicate and conflicting relationships
  • Ranks recommended changes
  • Never edits an input until you say yes

Install the Unify skill from the ZIP

Download the Unify ZIP from this page and unzip it. Then place the Unify folder in one of two locations: your personal skills folder at ~/.claude/skills/, or a project's .claude/skills/ folder if you want the skill available only inside that project.

If you use the Claude apps instead, upload the ZIP under Settings, then Capabilities, then Skills.

  • Personal install: ~/.claude/skills/
  • Project install: the .claude/skills/ folder inside your project
  • Claude apps: Settings → Capabilities → Skills, then upload the ZIP
~/.claude/skills/

Trigger Unify with natural requests

You do not need a special command. Ask in plain language and the agent will route the request to Unify when it matches the skill's trigger phrases. Trigger phrases from the skill include unify, harmonize, harmony check, are these in harmony, are these working against each other, do these conflict, where do these contradict, coherence check, unify LifeOS, is LifeOS coherent, are my systems fighting each other, alignment across components and bring these into unity.

Name the inputs when you can. If you name none and you are working with LifeOS, the skill analyzes the default LifeOS set.

  • Run Unify (default LifeOS scope)
  • Are the approval rules, the scheduling rule and the monitoring layer in harmony?
  • Unify these three strategy docs (paths given)
  • Do these two plans conflict?

Requirements and the default configuration

Unify needs the Bun runtime, because its CheckReport.ts script runs under bun. Full analysis also needs the other skills the SKILL.md refers to by name, which are FirstPrinciples, Superior and SystemsThinking.

For a run with no inputs named, Unify reads a config file at USER/CUSTOMIZATIONS/SKILLS/Unify/Config.md inside your LifeOS install. That file names the default input set and the summit the set is judged against. If the file is missing, the skill stops and points you to its setup notes rather than guessing. Named inputs do not need this file.

  • Bun runtime
  • FirstPrinciples, Superior and SystemsThinking installed for full analysis
  • LifeOS install with Config.md for no-input runs
~/.claude/LIFEOS/USER/CUSTOMIZATIONS/SKILLS/Unify/Config.md

Safety notes: scripts and network

Unify runs scripts. Its CheckReport.ts script checks a saved report against parts of the output contract that code can judge. The skill does not use the network.

Read-only does not mean the skill's conclusions are final. Whether a finding is real is a human judgment, so review each finding and recommendation before acting on it. The skill is also not designed for adversarial critique of a single plan.

  • Runs scripts: yes
  • Uses network: no
  • Read-only: recommends, does not edit
  • Findings need human review

Tips for useful results

Give Unify a clear set of inputs and, where possible, the sources behind each one. The skill is designed to cite both sides of a conflict, so inputs with clear aims produce clearer findings.

Use Unify for coherence across a set. For a feedback loop inside one system, the skill's notes route you to SystemsThinking. For attacking one plan, they route you to RedTeam. For editing personal goals, they route you to Telos.

  • Name the inputs and their source files
  • Expect vocabulary differences to show up as drift, which is often a rename rather than a real conflict
  • Use the report's ranked list as a starting point for human decisions

What people use it for

  • Checking whether approval rules, scheduling rules and monitoring layers in a LifeOS setup work against each other
  • Reviewing a set of strategy documents for coherence against a shared aim
  • Finding where two sets of rules for an agent contradict each other
  • Spotting duplicated or drifting goals across personal planning files

Frequently asked questions

How do I install the Unify agent skill?

Download the Unify ZIP from this page and unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. In the Claude apps, you can instead upload the ZIP under Settings, Capabilities, Skills.

What does Unify require to run?

It requires the Bun runtime because its CheckReport.ts script runs under bun. For full analysis, FirstPrinciples, Superior and SystemsThinking should be installed, and a no-input run needs a Config.md file in your LifeOS install.

Is Unify free to use?

This page does not state pricing or licence terms beyond the licence listed for the item, which is MIT. Check the repository's licence file for the full terms.

Does Unify edit my files?

No. Unify is read-only and only recommends changes. It does not edit an input until you approve the change.

Does Unify use the internet?

No, the skill does not use the network. It does run scripts, so review the scripts before use.

What is the difference between Unify and RedTeam or SystemsThinking?

Unify checks whether a set of inputs works in harmony. RedTeam is for attacking a single plan, and SystemsThinking covers feedback loops within one system, so the Unify skill routes those requests to them.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Unify suits owners of multi-part personal or team systems, builders of LifeOS-style agent setups, and anyone reviewing a set of planning documents for coherence. It is a poor fit if you want adversarial critique of one plan or a verified judgment without human review, so choose RedTeam or a human reviewer for those jobs.

Download the skill