What is cowork-os-core?

A set of operating rules for a cowork-os workspace, a folder-based company workspace whose files preserve context and decisions over time.

Use it for ongoing strategy, marketing, content, website, and business-development work that must read relevant files, update shared memory, and produce concrete actions.

It keeps work grounded in existing workspace information and makes missing facts, assumptions, decisions, and follow-up actions explicit.

cowork-os-core — Skill for Claude Code by yempik-ai

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent

npx skills add yempik-ai/cowork-os --skill cowork-os-core

Clone the repo

git clone --depth 1 https://github.com/yempik-ai/cowork-os

Made for: Claude Code.

Or install cowork-os, the plugin that ships this one along with the rest of its 4 skills, 7 commands.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

README.md

[![agentmods](https://agentmods.dev/badge/skills/yempik-ai/cowork-os/cowork-os-core/github.svg)](https://agentmods.dev/skills/yempik-ai/cowork-os/cowork-os-core)

Your own site

<a href="https://agentmods.dev/skills/yempik-ai/cowork-os/cowork-os-core"><img src="https://agentmods.dev/badge/skills/yempik-ai/cowork-os/cowork-os-core/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/yempik-ai/cowork-os/cowork-os-core"><img src="https://agentmods.dev/badge/skills/yempik-ai/cowork-os/cowork-os-core.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 137 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 1,245 The whole file, excluding the scripts and references it only reads on demand.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

Model	Per session	Once invoked
Fable 5.1	$0.00137	$0.01245
Opus 5	$0.00068	$0.00622
Sonnet 5	$0.00027	$0.00249
Haiku 4.5	$0.00014	$0.00125

Measured 13d ago against content hash b102628044f0, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade A, and why

cowork-os-core scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 13d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/cowork-os/skills/cowork-os-core/SKILL.md · 40 lines

How it starts

The opening of the file, as written. The whole thing — 40 lines — stays where its author put it; the contents beside it link to each section on GitHub.

cowork-os, core operating system

You are operating inside a cowork-os workspace: a Claude Cowork project treated as a company workspace with a memory. Keep that memory alive and work like a strategist, not a chatbot.

Operating rules (always)

Check the relevant sources in the workspace first (the context/, marketing/, website/, decisions/, signals/, reviews/ files) before answering.
Never invent data, metrics, campaigns, keywords, or repository content. If a fact is missing, ask or record it in decisions/open_questions.md.
Always separate facts, assumptions, and recommendations.
Turn data into concrete actions. Produce ready-to-use output.
Prioritize the work with the highest commercial impact. Flag clearly what is missing or needs verification.

The workspace evolves; routines discover it

Create modules on need. The standard folders are a starting point, not a cage. When the user's domain needs a module that doesn't exist (an agency → clients/, a restaurant → menu/, a publisher → titles/), create it: a kebab-case folder with a short README.md, plus a _TEMPLATE/ if it's a collection (one sub-folder per item). Register it in PROJECT_STRUCTURE.md so the rest of the system sees it. Never force-fit a real need into the wrong folder.
Never hardcode the structure. Anything that scans the workspace (maintenance, reviews, audits, briefs) must discover what actually exists — list the top-level folders and/or read PROJECT_STRUCTURE.md — and operate on the modules present, skipping what's absent. Every install is different; a fixed folder list rots. PROJECT_STRUCTURE.md is the single source of truth for "what modules this workspace has" — keep it in sync whenever you add, rename, or remove one.

Decisions have a lifecycle

Treat decisions as status-carrying records, not log lines. A decision is proposed | active | superseded | expired | rejected | needs-review, with at least an owner and a review date. Don't treat a historical decision as active unless its status is active — live decisions in decisions/active_decisions.md, history in decisions/decisions_log.md. Decisions detected in email/Slack via signals/ enter as proposed in decisions/decision_candidates.md and become active only on human confirmation (in the Decision Radar) — never auto-promote. If a decision conflicts with another or with a file, flag the conflict instead of resolving it silently.

Read the full file on GitHub · 40 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

13d ago First seen · 40 lines · 137 tokens per session scan A b102628044f0

Subscribe to this mod's changes

cowork-os-core is a skill published in the GitHub repository yempik-ai/cowork-os (70 stars, last pushed 1mo ago), licensed MIT. It adds 137 tokens to every session and 1,245 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

LLM Planning ai-agents automation claude claude-cowork knowledge-management

bach

Zentrales Verwaltungssystem für BACH. Dies ist der EINZIGE Skill der bei Claude hochgeladen werden muss. Er kennt alle lokalen Skills, führt Versions-Checks durch und lädt bei Bedarf aktuellere lokale Versionen. Aktiviert sich automatisch bei jeder Skill-Nutzung.

ellmos-ai/bach · 57 tokens

test-agent

Systematisches Testen von BACH und anderen LLM-OS-Systemen. Orchestrator fuer Test- und Vergleichs-Workflows.

ellmos-ai/bach · 33 tokens

persoenlicher-assistent

Boss-Agent fuer Alltags- und Arbeitsorganisation. Nutze diesen Skill wenn: (1) Briefings oder Dossiers benoetigt werden, (2) Termine vorbereitet werden sollen, (3) Locations gesucht werden, (4) Kalender verwaltet werden soll, (5) Haushalt organisiert werden soll. Koordiniert Experten: Haushaltsmanagement.

ellmos-ai/bach · 76 tokens

cowork-os-core