Cowork Plugin Engineer

Organization

microsoft

cowork-plugin-engineer

Use this skill whenever a user asks to create, import, repair, authenticate, package, validate, sideload, evaluate, upgrade, or troubleshoot a Microsoft Copilot Cowork plugin containing Agent Skills or remote Model Context Protocol (MCP) connectors.

Overview

Publisher	microsoft
Repository	`cat-agent-skills`
Skill name	`cowork-plugin-engineer`
Stars	95
Forks	115
Bundled files	16
License	MIT
Links	Skill source Repository

Markdown instructions
A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.
Works with any LLM
AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.
16 bundled files
Scripts, templates, and references the model can read while it works. Files are read-only and never executed.
Open source
Published by microsoft on GitHub. Read the source before you install it.

Installation

Install the Cowork Plugin Engineer AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

Open the app and go to Plugins → Skills.
Choose "Install from GitHub".
Paste the skill folder URL below and confirm.
https://github.com/microsoft/cat-agent-skills/tree/main/submissions/cowork-plugin-engineer
Enable the skill in any chat where you want it available.

Store

Installed

Skills

Add skills

Import one or more skills from a public GitHub folder or a local .skill or .zip archive, or SKILL.md file.

From GitHub URLFrom .skill, .zip, or SKILL.md

GitHub repo or folder

https://github.com/microsoft/cat-agent-skills/tree/main/submissions/cowork-plugin-engineer

Points to a single skill folder, or any folder that contains multiple skills. Public repos only.

CancelContinue

Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills

git clone --depth 1 https://github.com/microsoft/cat-agent-skills.git /tmp/cat-agent-skills
mkdir -p .claude/skills
cp -r /tmp/cat-agent-skills/submissions/cowork-plugin-engineer .claude/skills/cowork-plugin-engineer

Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Cowork Plugin Engineer in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Cowork Plugin Engineer on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Cowork Plugin Engineer is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

Loaded only when it is needed
The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.
Switch models mid-chat
Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Copilot Cowork Plugin Engineer

Build upload-ready Cowork plugins with deterministic validation. Treat the Microsoft 365 app package, its authentication configuration, and its deployment as separate gates.

Mandatory workflow

Read package-contract.md.
Search current Microsoft Learn documentation before choosing a manifest version or making rollout, preview, authentication, or licensing claims. For a manifest baseline upgrade, follow manifest-upgrade-instructions.md.
Inspect the source without modifying it. Detect:
- Native Microsoft 365 app package
- Claude, Cursor, or Agent Plugin package
- Skills-only folder
- Existing Agents Toolkit project
Choose the workflow:
- New project: use scripts/new_cowork_plugin_project.py.
- Open plugin import: follow import-and-normalization.md.
- Existing project: validate before editing.
- Existing ZIP: use scripts/test_cowork_plugin_package.py; never extract an untrusted archive without its path and expansion safeguards.
If an MCP connector is present:
- Discover tools through initialize, notifications/initialized, and tools/list; never invent tool names or schemas.
- Read authentication.md.
- Reject unresolved OAuth placeholders before producing a deployable ZIP.
When a source project exists, run scripts/test_cowork_plugin.py, then build through scripts/build_cowork_plugin.py. This uses Microsoft 365 Agents Toolkit (atk) for both packaging and package validation.
For a ZIP-only request, do not run project validation or rebuild it. Treat the ZIP accepted by scripts/test_cowork_plugin_package.py as the final package candidate.
Return the exact path of the built or successfully validated ZIP.
When evaluations are requested, read evaluations.md and use scripts/new_cowork_plugin_evals.py. Treat generated cases as drafts until every placeholder has an approved expected response.
Sideload or deploy only when the user explicitly asks. Keep tenant-wide deployment approval-gated.

Non-negotiable gates

Never modify an imported source tree; import from a temporary staging copy.
Never call a package deployable while OAuth referenceId contains a placeholder.
Never place secrets, tokens, client secrets, or API keys in a manifest, skill, script, or ZIP.
A registered skill folder contains exactly one root SKILL.md. Nested companion documents must use another filename such as REFERENCE.md or CAPABILITY.md.
Skill frontmatter name must be lowercase kebab-case and match its folder.
manifest.json must be at the ZIP root with no wrapper directory. Referenced assets may use nested paths, but must resolve within the package.
Preserve the manifest application ID across upgrades and increment the package version.
Do not use raw Compress-Archive as the final acceptance gate.
Do not fabricate a successful MCP connection from package validation; authentication and allowed-client configuration must also succeed.

Standard commands

Use python3 in the Cowork Linux runtime and on Linux/macOS. On Windows, replace python3 with py -3. Keep each command on one line so it works without shell-specific continuation syntax.

sh
# Validate a project without building.
python3 scripts/test_cowork_plugin.py --project-path <project>

# Validate an existing ZIP without trusting its contents.
python3 scripts/test_cowork_plugin_package.py --package-path <package.zip>

# Insert the Teams Developer Portal OAuth client registration ID.
python3 scripts/set_cowork_oauth_reference.py --project-path <project> --connector-id <connector-id> --oauth-configuration-id <generated-id>

# Build and validate the upload package.
python3 scripts/build_cowork_plugin.py --project-path <project>

# Generate a draft behavioral evaluation suite.
python3 scripts/new_cowork_plugin_evals.py --project-path <project>

Troubleshooting

Read troubleshooting.md and map the observed HTTP status, endpoint, and response body to a specific layer:

Package/manifest
OAuth configuration lookup
User sign-in or consent
MCP allowed-client policy
MCP protocol handshake
Tool invocation

Do not repeatedly retry configuration errors.

Dynamic Microsoft documentation

Topic	Search
Current Cowork package schema	`microsoft_docs_search(query="Build plugins for Copilot Cowork manifest validation rules")`
Connector authentication	`microsoft_docs_search(query="Cowork MCP OAuthPluginVault authentication configuration")`
Agents Toolkit commands	`microsoft_docs_search(query="Microsoft 365 Agents Toolkit CLI import openplugin package validate")`
Dataverse MCP setup	`microsoft_docs_search(query="Dataverse MCP remote endpoint allowed clients Entra app")`
Evaluation dataset schema	`microsoft_docs_search(query="Microsoft 365 Copilot Agent Evaluations CLI dataset schema test design")`

If Learn MCP tools are unavailable, use:

MCP tool	CLI equivalent
`microsoft_docs_search`	`npx --yes @microsoft/learn-cli search "<query>"`
`microsoft_docs_fetch`	`npx --yes @microsoft/learn-cli fetch "<url>"`
`microsoft_code_sample_search`	`npx --yes @microsoft/learn-cli code-search "<query>" --language <language>`

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Cowork Plugin Engineer AI skill do?

Why use Cowork Plugin Engineer on TypingMind?

Because you install it once and use it with any model. Cowork Plugin Engineer is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Cowork Plugin Engineer in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/microsoft/cat-agent-skills/tree/main/submissions/cowork-plugin-engineer. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Cowork Plugin Engineer?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Cowork Plugin Engineer?

As many as you like. As long as a model supports skills, you can use Cowork Plugin Engineer with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Cowork Plugin Engineer AI skill free?

Yes. It is published on GitHub by microsoft under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Discord Clawd

openclaw

OrganizationPopular

Use to talk to the Discord-backed OpenClaw agent/session; not for archive search.

Discrawl

Discord archive: search, sync freshness, DMs, summaries, TUI, repo/release work.

Gitcrawl

GitHub archive: issue/PR search, sync freshness, duplicate clusters, current GitHub handoff, and Gitcrawl repo work.

Graincrawl

Granola archive: search, sync freshness, notes, transcripts, panels, SQL counts, and Graincrawl repo work.

Overview

Installation

Install in TypingMind

Install in another agent

Use it in TypingMind

Works with any AI model

Skill instructions

Copilot Cowork Plugin Engineer

Mandatory workflow

Non-negotiable gates

Standard commands

Troubleshooting

Dynamic Microsoft documentation

References

Bundled files

Frequently asked questions

Related AI Skills

Set up your own AI workspace now