Mindket

One Skill to Rule Them All (task-observer)

Newversion 1 · updated
Star3927

Overview

Watches your AI sessions for corrections and repeated work, then suggests edits to your existing skills and flags candidates for new ones. You approve each change before anything is installed.

One Skill to Rule Them All (task-observer) is an open-source project on GitHub in the Productivity category, listed by Mindket. It is free. This is version 1, last updated on . Install it from GitHub as its README says. Get Mindket’s setup notes on your Mindket account, as a zip.

At a glance

You need
  • Node.js, for the skills command line tool
  • A CLAUDE.md file or your platform equivalent, to hold the activation instruction
Runs in
  • Agents and tools reached by the skills command line tool, such as Claude Code
  • The Claude apps: web, desktop, mobile and Cowork, through an uploaded skill bundle
  • Hermes and Openclaw setups, which users report working too
Start with
npx skills add rebelytics/one-skill-to-rule-them-all --skill task-observer

Platform

Claude CodeCoding agent
ClaudeAI chat

About

What it is

task-observer is a meta-skill that runs alongside your other skills and watches how they perform. When you do the same job manually again and again, it points that out as something a new skill could take over. It also notes corrections you make and suggests specific edits to skills you already have.

Who it suits

People who build and keep a library of AI skills and want it to improve from real use instead of staying frozen. It also suits anyone who keeps correcting the same thing in their agent output.

IndustryWho uses itHow they use itExample
ConsultingKnowledge workers with many repeatable tasksLet it flag recurring work as skill candidates and help build themA report format you keep fixing by hand
SoftwareDevelopers who write skills for coding agentsReview its suggested edits to tighten skill instructionsA skill that triggers in the wrong places
ContentWriters and editors with house style rulesTurn repeated style corrections into updates of a writing skillA tone fix you have made in several drafts
OperationsTeams that share a skill libraryUse its cross-cutting principles log to keep quality steady across skillsA rule that should apply to every new skill

Anyone with only a skill or two, or who does not mind writing skill updates by hand, can skip it.

Where to use it

It installs as a skill, so it runs wherever your skills load.

  • Agents and tools reached by the skills command line tool, such as Claude Code
  • The Claude apps: web, desktop, mobile and Cowork, through an uploaded skill bundle
  • Hermes and Openclaw setups, which users report working too
What it does for you
Use caseWhoWhat they get back
Find a new skillAnyone repeating manual workA flagged pattern and help turning it into a skill
Improve a skillSkill ownersSpecific update suggestions based on corrections and gaps
Raise the quality floorLibrary maintainersNew skills checked against shared cross-cutting principles

How to set it up

You need a place that loads skills, and Node.js if you use the terminal route. Installing is only half the job, because the skill must also be switched on.

  • Node.js, for the skills command line tool
  • A CLAUDE.md file or your platform equivalent, to hold the activation instruction
  1. Run the terminal command below. In the Claude apps, upload the .skill file from the newest release in the settings instead.
  2. Copy the activation instruction found in references/environments.md into your CLAUDE.md, or the matching file on your platform.
  3. Check that it runs, using the check described in the project documentation.
npx skills add rebelytics/one-skill-to-rule-them-all --skill task-observer

Adding the -g option installs it for all your projects instead of one.

How to use it

Work as you normally do. The skill runs in the background of each session and writes down what it notices. You read the observations and decide which ones become changes.

What it produces
ToolWhat it does
Observation logOne entry per observation, with the skills it touches and a proposed fix
Cross-cutting principles logHolds patterns that apply to many skills, and new skills are checked against them
Staged copiesWhere scheduled reviews put changes that need no decision from you

Example

You correct an agent three times because its weekly summary keeps the wrong tone. The observer logs that as a signal, names the summary skill as affected and proposes a specific edit. You approve it, then install the updated bundle.

For newcomers and experts

If you are new, install it, switch it on, then work as usual for a few sessions. Read the observations it writes and approve one small change to see the cycle.

If you know the field, the point is the gap between project memory and a reusable skill. Project memory stores a fact for one project and applies it as an instruction. An observation here is a proposal that reaches every project loading the skill, and a human review sits in between.

  1. You work and correct the agent
  2. Observer logs corrections and gaps
  3. It suggests skill changes
  4. You review and approve
  5. You install the updated bundle

Good to know

  • License: CC-BY-4.0.
  • Your installed skills are left as they are, and changes take effect only once you install the new bundle.
  • Scheduled reviews only apply no-decision changes to staged copies.
  • The skill description alone tends to under-trigger, so the activation instruction is needed.
  • The author reports more than 2,300 logged observations across their own 87 skills.
  • The project credits over 100 contributors and reports over 17,000 installs on the skills.sh listing.

README and files

README
# One Skill to Rule Them All (task-observer)

Watches your AI sessions for corrections and repeated work, then suggests edits to your existing skills and flags candidates for new ones. You approve each change before anything is installed.

Made by rebelytics on GitHub. These notes are Mindket's, written from its README; the project's own files and license are in its repository.

## About

What it is: task-observer is a meta-skill that runs alongside your other skills and watches how they perform. When you do the same job manually again and again, it points that out as something a new skill could take over. It also notes corrections you make and suggests specific edits to skills you already have.

Who it suits: People who build and keep a library of AI skills and want it to improve from real use instead of staying frozen. It also suits anyone who keeps correcting the same thing in their agent output.

```table
Industry | Who uses it | How they use it | Example
Consulting | Knowledge workers with many repeatable tasks | Let it flag recurring work as skill candidates and help build them | A report format you keep fixing by hand
Software | Developers who write skills for coding agents | Review its suggested edits to tighten skill instructions | A skill that triggers in the wrong places
Content | Writers and editors with house style rules | Turn repeated style corrections into updates of a writing skill | A tone fix you have made in several drafts
Operations | Teams that share a skill library | Use its cross-cutting principles log to keep quality steady across skills | A rule that should apply to every new skill
```

Anyone with only a skill or two, or who does not mind writing skill updates by hand, can skip it.

Where to use it: It installs as a skill, so it runs wherever your skills load.

- Agents and tools reached by the skills command line tool, such as Claude Code
- The Claude apps: web, desktop, mobile and Cowork, through an uploaded skill bundle
- Hermes and Openclaw setups, which users report working too

```table
Title: What it does for you
Use case | Who | What they get back
Find a new skill | Anyone repeating manual work | A flagged pattern and help turning it into a skill
Improve a skill | Skill owners | Specific update suggestions based on corrections and gaps
Raise the quality floor | Library maintainers | New skills checked against shared cross-cutting principles
```

How to set it up: You need a place that loads skills, and Node.js if you use the terminal route. Installing is only half the job, because the skill must also be switched on.

- Node.js, for the skills command line tool
- A CLAUDE.md file or your platform equivalent, to hold the activation instruction

1. Run the terminal command below. In the Claude apps, upload the .skill file from the newest release in the settings instead.
2. Copy the activation instruction found in references/environments.md into your CLAUDE.md, or the matching file on your platform.
3. Check that it runs, using the check described in the project documentation.

```bash
npx skills add rebelytics/one-skill-to-rule-them-all --skill task-observer
```

Adding the -g option installs it for all your projects instead of one.

How to use it: Work as you normally do. The skill runs in the background of each session and writes down what it notices. You read the observations and decide which ones become changes.

```table
Title: What it produces
Tool | What it does
Observation log | One entry per observation, with the skills it touches and a proposed fix
Cross-cutting principles log | Holds patterns that apply to many skills, and new skills are checked against them
Staged copies | Where scheduled reviews put changes that need no decision from you
```

Example: You correct an agent three times because its weekly summary keeps the wrong tone. The observer logs that as a signal, names the summary skill as affected and proposes a specific edit. You approve it, then install the updated bundle.

For newcomers and experts: If you are new, install it, switch it on, then work as usual for a few sessions. Read the observations it writes and approve one small change to see the cycle.

If you know the field, the point is the gap between project memory and a reusable skill. Project memory stores a fact for one project and applies it as an instruction. An observation here is a proposal that reaches every project loading the skill, and a human review sits in between.

```flow
You work and correct the agent
Observer logs corrections and gaps
It suggests skill changes
You review and approve
You install the updated bundle
```

Good to know:

- License: CC-BY-4.0.
- Your installed skills are left as they are, and changes take effect only once you install the new bundle.
- Scheduled reviews only apply no-decision changes to staged copies.
- The skill description alone tends to under-trigger, so the activation instruction is needed.
- The author reports more than 2,300 logged observations across their own 87 skills.
- The project credits over 100 contributors and reports over 17,000 installs on the skills.sh listing.

## On GitHub, 16 November 2026

- Stars: 3,924
- Forks: 390
- Watching: 13
- Open issues and pull requests: 33
- Language: Markdown
- License: CC-BY-4.0
- Topics: agent-skills, ai-skills, augmented-expertise, awesome-skills, claude, claude-code, claude-code-skill, claude-code-skills, claude-cowork, claude-skills, meta-skill, rebelytics

## Links

- Code: https://github.com/rebelytics/one-skill-to-rule-them-all
- Website: https://www.rebelytics.com/task-observer/
- Releases: https://github.com/rebelytics/one-skill-to-rule-them-all/releases
- Issues: https://github.com/rebelytics/one-skill-to-rule-them-all/issues
- Star history: https://www.star-history.com/#rebelytics/one-skill-to-rule-them-all&Date

## Get the code

```
git clone https://github.com/rebelytics/one-skill-to-rule-them-all.git
```

Then follow the setup in its README. Nothing in this kit runs by itself.
Files
README.md  6 KB

1 file, 6 KB, downloaded as a zip.

What’s new

  • Version 1 · 2026-11-16

Compare with similar

Listed GitHub projects in the same category that share the most topics with this one.

ProductStarsActivityLanguageLicenseLast change
One Skill to Rule Them All (task-observer) This page3,927ActiveMarkdownCC-BY-4.0
AIEVShares claude-skills · claude · claude-code124ActiveTypeScriptMIT
Frontend SlidesShares claude · claude-code32,301QuietJavaScriptMIT
Numbers as Mindket last read them, from 2026-11-06 to 2026-11-16. Activity is read from each project’s last change on GitHub: Active within 30 days, Quiet within 180, Stale after that.

Reviews

No ratings yet

Write a review

No reviews yet.

What people say

Posts and comments that name this project on Hacker News, Bluesky, X, Reddit, Facebook and YouTube, found by a weekly search and on Trendshift. Mindket doesn’t write or check them. Open one to read it where it was posted.

Searched

More from Mindket

See all from Mindket

all-mpnet-base-v2

sentence-transformers · Kit

Turns sentences and paragraphs into 768-number vectors so you can search, cluster and compare text by meaning. It suits developers and data teams working in En…

New

multilingual-e5-base

intfloat · Kit

Turns text in about 100 languages into vectors so you can compare meaning across them, for search and matching. Made for developers who build multilingual retr…

New

sprite-gen

aldegad · Skill

Turns one base drawing of a character into game-ready sprites, either an atlas with a manifest or transparent motion loops, for game developers working in Code…

New4097

Similar projects

See all in Productivity

pstack

michael-denyer · Skill

Michael Denyer’s port of Lauren Tan’s pstack skill stack to Claude Code, Codex, Pi, GitHub Copilot, Muse and other agents. Tell its poteto-mode skill your goal…

New4132

marketingskills

coreyhaines31 · Skill

A set of marketing skills for coding agents, covering conversion, copywriting, SEO, analytics and growth. Built for technical marketers and founders who use Cl…

New60.6k

agents by wshobson

wshobson · Skill

Offers a catalog of plugins that bundle agents, skills and commands for AI coding tools including Claude Code, Codex CLI, Cursor and GitHub Copilot, kept as on…

New42.6k