richfrem/agent-plugins-skills / plugins/agent-extension-authoring/.apm/skills/packaging-copilot-cowork-plugins
packaging-copilot-cowork-plugins
VerifiedCreates, validates and prepares Microsoft 365 Copilot Cowork plugin App Packages containing Agent Skills and optional remote MCP connectors. Use when scaffolding, converting, validating, provisioning, packaging or sharing a Cowork plugin. Not for APM package creation or installation (use packaging-with-apm).
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 4.2 KB |
| evals/evals.json | config | 1.6 KB |
| evals/task-success.json | config | 3.1 KB |
| packaging-copilot-cowork-plugins-flow.mmd | other | 651 B |
| references/acceptance-criteria.md | markdown | 3.1 KB |
| references/cowork-capabilities-and-limits.md | markdown | 3.7 KB |
| references/cowork-deterministic-vs-generative.md | markdown | 3.5 KB |
| references/cowork-package-lifecycle.md | markdown | 13.9 KB |
| references/cowork-sandbox-findings.md | markdown | 7.0 KB |
| references/cowork-test-playbook.md | markdown | 3.6 KB |
| references/skill-install-routes.md | markdown | 4.7 KB |
| scripts/bundle_skill_companions.py | script | 19.8 KB |
| scripts/cowork_content_checks.py | script | 9.9 KB |
| scripts/cowork_markdown_utils.py | script | 4.8 KB |
| scripts/cowork_package_constants.py | script | 5.2 KB |
| scripts/package_cowork_plugin.py | script | 21.0 KB |
| scripts/scaffold_cowork_package.py | script | 13.4 KB |
| tests/test_bundle_and_scaffold.py | script | 10.9 KB |
| tests/test_content_checks.py | script | 8.9 KB |
| tests/test_package_cowork_plugin.py | script | 9.2 KB |
references/cowork-capabilities-and-limits.md
View on GitHub# Cowork Capabilities and Limits
What a skill can ask Cowork to do, and what it cannot rely on. Each row carries an evidence
label so the table is not mistaken for a guarantee.
- **Observed**: seen in a Cowork reply or a local test on 2026-11-18.
- **Reported**: stated by a Cowork user from their own use, not re-tested here.
- **Not tested**: designed, but no Cowork result collected.
Details and the reasoning behind each row are in
[the sandbox findings](cowork-sandbox-findings.md).
## Contents
- [What runs](#what-runs)
- [What does not](#what-does-not)
- [Skill and file limits](#skill-and-file-limits)
- [How to read this table](#how-to-read-this-table)
## What runs
| Capability | Status | Notes |
|---|---|---|
| Python 3.12 scripts | Observed | Only with libraries already installed: pdfplumber, pypdf, pypdfium2, openpyxl. Scripts ran from a skill in earlier sessions. |
| pandoc 3.9, LibreOffice 25.8 (`soffice`), pdftotext, tesseract (English) | Observed | Called from a script or shell command. |
| JavaScript and TypeScript (`.js`) | Reported | Limited libraries; no package installs. |
| A static Linux x86_64 binary (for example Rust built for `x86_64-unknown-linux-musl`) | Observed | Only from `/tmp` after `chmod +x`, and the file has to be attached to the chat. |
| Writing files back | Observed | Output to the working folder, then saved to the task output folder, which syncs to OneDrive. |
| Reading a skill's own files with the file viewer | Observed | Works for `SKILL.md`, scripts, assets and binaries. |
| Generating images from a program (PNG written by a binary) | Observed | The agent showed and saved the image. |
## What does not
| Limit | Status | What to do instead |
|---|---|---|
| Compiling Rust (or other compiled code) | Reported | Build off Cowork, test under WSL, ship the result. |
| Installing packages (`pip install`, `npm install`) | Observed | There is no network. Use what is present or ship it inside the skill. |
| PyMuPDF (`fitz`) | Observed | Use pdfplumber and pypdf. |
| Executable flag in the workspace | Observed | `chmod +x` there reports success but does nothing. Run from `/tmp`. |
| Skill files on disk in the command sandbox | Observed | Often absent. Reading them with the viewer and writing a copy into `/tmp` is untested (the agent said it could not copy a skill binary); otherwise ask for an attachment. |
| Agent copying a skill binary into the session itself | Not tested | The agent said it could not. The latest test skills ask it to try every tool and report. |
| Interactive programs | Observed | Cowork sees command output and files after a command ends. Use a deterministic mode with arguments. |
| A web page running a Linux binary | Observed | Compile the shared logic to WebAssembly and embed the `.wasm`. |
| Running an unknown binary without approval | Observed | Cowork declines; it ran the program once told it is the user's own. |
| Same-name re-upload refreshing a skill | Observed | The old files kept running. Use a new name per build. |
## Skill and file limits
- Upload error seen: a skill may have `SKILL.md` plus at most 20 other files.
- Packager limits in this repo: 20 companion files per skill, 5 MB per file, 10 MB total.
Microsoft Learn documents larger limits for a plain skill upload (see
[install routes](skill-install-routes.md)); use the stricter ones for portability.
- Frontmatter must be valid YAML. An unquoted `run: Python` in a description made Cowork show
the skill with no description.
- Plan for one architecture: Linux x86_64.
## How to read this table
Treat Observed rows as true for the sessions tested, not for every session. Run the small
test skills described in [the test playbook](cowork-test-playbook.md) before building a
skill that depends on a row, and record any change here.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |