Grimoire 013 · Claude Cowork 101 · Delegation Without The Terminal
13 modules · 7 blueprints · 14 prompt templates · 8 cited sources · Free, no gate
Grimoire 013 · Claude Cowork 101 · June 2026

Claude Cowork 101: Delegation Without The Terminal

A field manual for workers who live in files, folders, notes, meetings, PDFs, inboxes, and client messages. The goal is not to make everyone a developer. The goal is to make messy office work delegable without surrendering judgment.

13 modules
7 blueprints
14 prompt templates
5 approval levels
8 cited sources
WEEK
Speedrun: First Week With Cowork
Build trust in layers before expanding access.
  1. Day 1 - Read-only summaries. Give one folder or small document set. Ask for a summary and source list. No edits.
  2. Day 2 - Briefs from bounded sources. Ask for a one-page brief with risks, open questions, and next actions.
  3. Day 3 - Comparisons. Compare two documents or versions. Ask for material differences only.
  4. Day 4 - Drafting. Prepare a client update, memo, checklist, or meeting brief. Do not send.
  5. Day 5 - Folder map. Ask for a proposed organization plan. Still no moving files.
  6. Day 6 - Local artifact creation. Allow creation of a new draft file in an approved folder.
  7. Day 7 - Review the workflow. Keep the tasks that saved time. Remove the ones that created confusion.
First-week rule: Cowork should produce visible artifacts before it changes existing work.
WHY
The Missing Middle
Between chatbot and coding agent sits the office worker.

Claude Code makes sense to developers because developers already live in repositories, terminals, diffs, tests, commits, and pull requests. Most workers live somewhere else.

They live in folders, PDFs, spreadsheets, inboxes, notes, meetings, screenshots, client messages, and half-finished documents. Cowork matters because useful AI cannot stay trapped inside developer workflows.

Thesis

Cowork is not "AI for non-technical people." It is delegated office execution with boundaries.

The worker still owns judgment. Cowork helps carry the search, comparison, drafting, formatting, and packaging that normally sit invisibly inside office work.

1
bounded source set
1
visible artifact
0
external sends without approval
M1
Cowork vs Chat
Chat answers. Cowork works from selected material and leaves artifacts.
ModeWhat it doesReview surfaceRisk
ChatAnswers, drafts, explains, summarizes pasted contentConversationContext may be incomplete or stale
CoworkReads selected files, compares, drafts, creates artifacts, proposes changesFile, table, brief, source list, change logMay act on wrong source or hidden assumption
Claude CodeWorks in codebases, tests, commits, branches, PRsDiff, test output, commitCode changes may exceed scope
OpenClawPersistent assistant shell reachable through channels and webhooksSkill output, channel message, completion reportPermission creep if not governed

The point is not which one is better. The point is which body fits the work. A worker who understands the difference rarely hands the wrong task to the wrong tool.

Decision shortcut: if the job starts with a folder, a file, a PDF, or an inbox — that is Cowork territory. If it starts with a question or a blank page — that is chat. If it starts with a repo — that is Claude Code.
Cowork advantage

The artifact is the review surface

When Cowork works on files, you can see exactly what it read, what it produced, and what it assumed. The artifact is inspectable in a way that a chat response is not.

Cowork limit

Still needs source discipline

Cowork is only as reliable as the source set you give it. Vague access (“use my files”) produces vague work. Bounded access produces inspectable work.

M2
The Delegation Contract
A Cowork task needs five parts.
Cowork Delegation Contract
Task:
[What you want done]

Source material:
[Files, folders, notes, links, or context]

Output:
[Exact format you want]

Boundaries:
[What not to touch, what requires approval]

Check:
[How you want it to verify the result]

This is not bureaucracy. It is how the worker stops the agent from guessing where judgment is required. Anthropic ↗

Worker Rule

If the task has consequences, write the boundaries first. Most bad Cowork outcomes start because the assistant was asked to solve an undefined problem with undefined permissions.

M3
Source Material
The source set decides the quality of the result.

Cowork is only as reliable as the material it is allowed to inspect. "Use the project files" is weaker than "Use only the files in /Project Notes/June and ignore archive folders."

Source typeGood instructionBad instruction
FolderUse only this folder and list every file used.Look around and find what matters.
PDFsSummarize these three PDFs and cite page or section names where possible.Summarize the documents.
NotesUse notes dated after May 1 and flag contradictions with older notes.Use my notes.
Email exportExtract tasks only. Do not draft or send replies.Handle this thread.
Source Discipline Prompt
Before drafting, list:
- every source you will use
- every source you will ignore
- anything missing
- anything that looks outdated or contradictory
M4
Output Artifacts
The output must be inspectable.

Cowork should produce something the worker can review: a brief, table, checklist, draft, folder map, source list, task list, or change proposal.

Good artifact

One-page project brief

Summary, timeline, risks, open questions, next actions, source list.

Weak artifact

"Here is what I found"

Too vague to approve, repeat, or hand off.

Good artifact

Comparison table

Difference, impact, recommendation, confidence, source.

Weak artifact

Long narrative summary

Hard to scan, hard to verify, hard to turn into action.

M5
Folder Work
A folder is project memory. Treat it carefully.

Never begin folder work by moving files. Begin with a map. Anthropic ↗ frames this as read-before-write: let the system observe before it acts.

Folder Map Prompt
Read this folder and propose an organization plan.

Do not rename, move, delete, upload, or edit anything yet.

Return:
- current structure summary
- file categories
- duplicate or outdated files
- proposed folder structure
- files you are unsure about
- risks before changing anything
Folder rule: proposal first, approval second, movement third.
M6
Document Work
Summaries are only the first layer.

Document work becomes valuable when the assistant can extract decisions, risks, contradictions, and next actions instead of merely compressing text.

Document Brief Prompt
Read the selected documents.

Output:
- executive summary
- key facts
- decisions already made
- unresolved questions
- contradictions
- risks
- recommended next action
- source list

Do not invent missing facts. Mark uncertainty clearly.
Document typeBest outputExtra check
ProposalScope, cost, assumptions, exclusionsFind vague obligations
ContractNon-legal summary and risk flagsRequire human/legal review
ReportFindings, evidence, recommendationSeparate data from interpretation
SOPStep list, owner, trigger, failure modeFlag missing decision points
M7
Meeting Work
Meetings fail when decisions do not become tasks.

The useful output of a meeting is not a transcript. It is a decision record, task list, and risk register.

Meeting Debrief Prompt
Turn these meeting notes into:
- decisions
- tasks
- owners
- deadlines
- blockers
- open questions
- follow-up message draft

If owner or deadline is missing, mark it as missing.
Do not invent names or dates.
Operator standard: if a meeting produces no owner, no due date, and no next action, it produced conversation, not execution.
M8
Spreadsheet Work
Use Cowork for interpretation and structure, not blind financial authority.

Spreadsheet work needs extra caution because a small formula mistake can create large business consequences. Use Cowork to inspect, summarize, and flag issues before allowing edits.

Spreadsheet typeSafe Cowork roleRequires human review before
Data export / listClassify, count, summarize, find duplicatesAny structural edit
Project trackerExtract tasks, owners, deadlines, blockersChanging status fields
Budget or forecastRead totals, flag anomalies, note missing rowsAny formula or value edit
Payroll / invoicingVerify structure and flag missing data onlyAll of it — escalate to owner
Client reportingCheck numbers against source data, note discrepanciesAny change to client-facing figures
Spreadsheet Inspection Prompt
Inspect this spreadsheet.

Return:
- sheet purpose
- columns and meanings
- obvious data quality issues
- missing values
- suspicious formulas or hardcoded numbers where a formula is expected
- summary statistics
- questions before editing

Do not modify the file.
Do not recalculate or reinterpret formulas — report what is there.
Never skip review for payroll, taxes, financial forecasts, invoices, legal calculations, or anything client-billable.
Formula risk: Cowork can read formula text but cannot always predict downstream effects in complex multi-sheet workbooks. Flag suspicious dependencies and let the owner verify before changing anything.
M9
Inbox Work
Inbox delegation should start with extraction, not sending.

The inbox contains commitments, private context, money, legal issues, client expectations, and emotional landmines. Do not begin by letting an agent "handle email." Begin with triage.

Inbox Triage Prompt
Review the selected messages.

Extract only:
- urgent items
- needs reply
- waiting on someone else
- FYI/archive
- deadlines
- commitments I made

Do not draft, send, archive, delete, unsubscribe, or mark anything read unless I approve.
StageAllowedRequires approval
TriageClassify, extract tasks, summarizeChanging message state
DraftPrepare reply draftsSending or reply-all
CleanupRecommend filters or foldersUnsubscribe, delete, archive
M10
Approval Levels
Do not grant Level 5 trust for Level 1 performance.
LevelPermissionExamplesDefault for beginners?
1Read-onlySummarize, classify, compare, extractYes
2DraftingCreate briefs, memos, checklists, reply draftsYes
3Local editsCreate or edit files in approved foldersAfter review
4External actionsSend, publish, create tickets, update systemsNo
5Sensitive actionsMoney, credentials, deletion, legal, productionNo
Approval Gate
Before any Level 4 or Level 5 action, stop and ask:
- What action will be taken?
- What source supports it?
- What could go wrong?
- How can it be reversed?
- Do you approve?
M11
QA and Review
Visible work is not enough. It must be checkable.

Every Cowork output should include a review trail. The worker should not need to guess whether the assistant used the right files or invented missing details.

Cowork Completion Report
Completed:
- [artifact created]

Sources used:
- [file/link/note]

Not used:
- [ignored source and why]

Assumptions:
- [inferences]

Uncertainties:
- [unknowns]

Needs approval:
- [decision/action]
Review Rule

Ask for uncertainty explicitly. If you do not create a place for uncertainty, many assistants will bury it under fluent prose.

M12
7 Cowork Blueprints
Safe, useful starting workflows.
BP1

Project Brief

Selected notes in, one-page brief out: summary, timeline, risks, questions, next actions.

BP2

Folder Map

Approved folder in, proposed structure out. No movement until approved.

BP3

Proposal Comparison

Two proposals in, material differences table out: scope, cost, risk, recommendation.

BP4

Meeting Action List

Notes in, decisions, owners, deadlines, blockers, and follow-up draft out.

BP5

Client Update Draft

Source notes in, client-ready draft out. Do not send.

BP6

Inbox Task Extraction

Selected emails in, commitments and deadlines out. No sending or archiving.

BP7

SOP Builder

Messy process notes in, trigger, owner, steps, checks, and failure modes out.

Blueprint prompt pack
BP1 - Project Brief
Create a project brief from the selected files.

Use only the selected files.
Output:
- summary
- timeline
- decisions
- risks
- open questions
- next actions
- source list

Do not move, rename, delete, send, or upload anything.
BP3 - Proposal Comparison
Compare these two proposals.

Return a table:
- item
- Proposal A
- Proposal B
- material difference
- operational impact
- question to ask before choosing
BP7 - SOP Builder
Turn these rough notes into an SOP.

Include:
- trigger
- owner
- inputs
- steps
- checks
- failure modes
- escalation point
- example completed output
M13
Assessment: Can This Be Delegated?
A task is ready for Cowork only when the answers are concrete.

1. What exact artifact should Cowork produce?

2. Which files, folders, or notes are allowed sources?

3. Which sources should be ignored?

4. What actions are forbidden?

5. What requires approval?

6. How will the output be checked?

7. What uncertainty should be flagged?

8. What would make this task unsafe to automate?

9. Where should the artifact be saved?

10. What is the next human decision?

Pass/fail standard

Pass if the task has bounded sources, visible output, explicit forbidden actions, and a review method. Fail if the prompt says "handle," "clean up," "fix," or "organize" without defining scope and approval.

PROMPT
Prompt Library
Quick-reference prompts for using Cowork during real work.

This library points to the major prompt blocks throughout the manual and fills in the missing field prompts so the page works as a reference desk.

PromptUse it whenJump
Cowork Delegation ContractYou need to define task, sources, output, boundaries, and checks.M2 - Contract
Source Discipline PromptYou need the assistant to declare what it will use before drafting.M3 - Sources
Artifact TypesYou need to choose the right output shape before delegating.M4 - Artifacts
Folder Map PromptYou need a folder organization proposal without moving files.M5 - Folder
Document Brief PromptYou need decisions, risks, contradictions, and next actions from documents.M6 - Documents
Meeting Brief PromptYou need meeting notes turned into decisions, owners, deadlines, and blockers.M7 - Meetings
Inbox Triage PromptYou need tasks and commitments extracted from selected messages.M9 - Inbox
Approval PromptYou need to stop before Level 4 or Level 5 actions.M10 - Approval
QA Review PromptYou need source trail, assumptions, uncertainties, and review points.M11 - QA
Blueprint Prompt PackYou need reusable prompts for project briefs, comparisons, and SOPs.M12 - Blueprints

Approval Prompt

Approval Prompt
Before taking this action, stop and prepare an approval request.

Include:
- exact action
- destination or system affected
- source evidence
- risk if wrong
- rollback path
- what you need me to approve

Do not act until approval is explicit.

QA Review Prompt

QA Review Prompt
Review this Cowork output before I use it.

Check:
- sources used
- missing sources
- unsupported claims
- assumptions
- uncertainty
- formatting issues
- actions that require approval
- next human decision
GLO
Glossary
Terms used in this manual.
TermMeaning
CoworkAn AI delegation pattern for office/file work through a friendly interface rather than a terminal.
ArtifactThe visible output of work: brief, table, checklist, draft, file map, or report.
Source setThe bounded material the assistant is allowed to use.
Approval gateA required human decision before consequential action.
Visible workOutput plus source trail, assumptions, uncertainties, and next decision.
REF
Source Anchors
This draft treats Cowork as a work pattern and Claude product direction. Re-verify product-specific availability before publishing.
01
Claude Code overview - agentic surfaces, file editing, command execution, desktop, web, and IDE context.
Link
02
Claude Code common workflows - task planning, codebase exploration, subagents, worktrees, and scripting patterns.
Link
03
Claude Code skills - packaging repeatable workflows for teams: triggers, tool access, and sharing.
Link
04
Claude Code memory and CLAUDE.md - persistent project instructions, auto-memory, and session context.
Link
05
Anthropic Engineering - Building Effective Agents: composable patterns, tool design, human-in-the-loop, and safe delegation.
Link
06
Model Context Protocol documentation - open standard for connecting AI applications to external tools and data sources.
Link
07
Anthropic usage policy - permitted and prohibited uses, human oversight requirements, and operator responsibilities.
Link
08
Aaron Suarez Working Library - grimoire positioning: practical field manuals, free and ungated.
Link
Publication note: Cowork is an evolving product. Before final release, verify current product naming, supported file types, folder permissions, platform availability, and any enterprise controls against official documentation.
Related Grimoires
012
Claude + OpenClaw 101
Persistent assistants, channels, memory, skills, and operator-grade execution.
007
Agentic AI 101
Agent architectures, guardrails, evals, and the gap between demos and dependable workflows.
006
Automations 101
25 production-grade automation patterns — idempotency, retries, DLQs, runbooks.
014
Remote Work 101
The front gate — 29 chapters on remote careers, systems, survival, and the Philippine reality.
Resume where you left off?Jump back