Guide · Agent skill
How to use the Wrap-up Agent Closeout skill (wrap-up)
Updated · written with AI from the project's own documentation and checked against our catalogue data
Wrap-up is an agent skill for closing out a Claude Managed Agent that you have already built and launched. It lists what you now control, regenerates a single HTML overview page, and suggests your next one or two upgrades. This guide explains what the skill does, how to install it, how to trigger it, and what it needs before it will work.
Wrap-up Agent Closeoutalirezarezvani/claude-skills · MIT Download the skill Quick start
- 1Install the skill. Unzip the wrap-up ZIP into ~/.claude/skills/ or a project's .claude/skills/ folder, or upload it under Settings → Capabilities → Skills in the Claude apps.
- 2Check your build sheet. Confirm that ./my-agent/build-sheet.json and ./my-agent/goal.json exist and reflect the current agent setup.
- 3Ask for a wrap-up. Say something like "wrap up" or "what do I own now" so the agent matches the request to the skill.
- 4Review the inventory. Read the tables of agent, environment, session, memory, outcome and deployment items the script prints.
- 5Open the overview page. Open ./my-agent/agent-overview.html in a browser and share the single file if needed.
- 6Choose your next upgrade. Start with the top-ranked suggestion from the upgrade suggester and follow the mechanism it names.
What is an agent skill, and what does wrap-up do?
An agent skill is a folder of instructions and scripts that an AI agent can load and follow when a request matches the skill's description. The wrap-up skill is the final step in the agent-launcher workflow. It is meant for the moment after an agent is live, when you want to know exactly what you own.
When run, the skill inventories the agent's parts (agent, environment, session, memory, outcome and deployment) and lists the phases completed. It then writes a self-contained ./my-agent/agent-overview.html file and ranks the next moves from deferred items and standard hardening steps.
- Reads ./my-agent/build-sheet.json and ./my-agent/goal.json
- Lists every primitive the owner controls, including model, tools, environment, memory and deployment schedule
- Regenerates agent-overview.html with inline CSS and no external assets
- Suggests the top 1-2 next upgrades, each with the exact mechanism to use
Install the wrap-up skill
Download the ZIP for the wrap-up skill from the directory page. Then choose one of two locations. For a personal install available across projects, unzip it into ~/.claude/skills/. For a single project, unzip it into the project's .claude/skills/ folder.
If you use the Claude apps instead, open Settings, go to Capabilities, and upload the skill under Skills.
- Personal: unzip into ~/.claude/skills/
- Project only: unzip into .claude/skills/ inside the project
- Claude apps: Settings → Capabilities → Skills, then upload
Prepare the files the skill expects
Wrap-up does not create your agent. It reads the records left by the wider agent-launcher workflow. Before running it, check that ./my-agent/ contains build-sheet.json and goal.json. The recap is only as accurate as the build sheet, so update it first if you have changed anything.
The skill is told never to claim a primitive that was not created, so an out-of-date sheet will produce an out-of-date recap.
- ./my-agent/build-sheet.json must exist and be current
- ./my-agent/goal.json must exist
- You need Python 3; the scripts use only the standard library and no third-party packages
Trigger the skill with natural requests
You do not need to type a command. Ask in plain language and the agent should match the request to the skill. The skill's description lists these phrases as triggers: "wrap up", "close this out", "what do I own now", "give me the summary" and "recap the agent".
When the agent runs the skill, it works through three scripts in order: the inventory, the overview page and the upgrade suggester.
- "Wrap up the agent."
- "What do I own now?"
- "Give me the summary of this agent."
- "Recap the agent and suggest what to do next."
python3 scripts/primitives_inventory.py --sheet ./my-agent/build-sheet.json --goal ./my-agent/goal.json
python3 scripts/overview_page.py --sheet ./my-agent/build-sheet.json --out-dir ./my-agent --status live --loop-shape cron-loop --last-verdict satisfied
python3 scripts/upgrade_suggester.py --sheet ./my-agent/build-sheet.json --top 2Read the outputs
The inventory prints tables for each primitive and the phases completed. The overview page is saved as ./my-agent/agent-overview.html and can be shared as a single file. The upgrade suggester lists the highest-ranked next moves.
Deferred items are ranked with v1 before v2 and real-integration work first. Standing hardening steps include tightening networking, pinning the agent version and nesting an outcome.
- Inventory: tables of owned primitives and completed phases
- Overview: ./my-agent/agent-overview.html, self-contained and theme-aware
- Upgrades: top-ranked deferrals plus standing hardening steps
Safety notes: scripts and network
Wrap-up runs local Python scripts. The source states that it runs scripts and does not use the network. Review the scripts in the scripts/ folder if you want to check exactly what they do before running them.
Because the overview page is a single file with no external assets, sharing it does not require sending it to any other service.
- Runs scripts: yes (Python 3, standard library only)
- Uses network: no
- Overview page has no external assets
What people use it for
- Recapping what is live after launching a Claude Managed Agent
- Creating a shareable single-file overview of an agent setup for a team
- Getting a ranked list of next improvements from earlier deferred items
- Refreshing the overview page after changing the build sheet
Frequently asked questions
Is the wrap-up skill free to use?
The source is released under the MIT licence, which is stated on the directory page. Check the licence file in the download for the full terms. The skill itself does not state any price.
What do I need before I can use wrap-up?
You need Python 3, an existing ./my-agent/build-sheet.json and goal.json produced by the agent-launcher workflow, and a Claude Managed Agent setup to summarise. The skill is not useful without those files.
Where do I install the wrap-up skill?
Unzip the ZIP into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder for one project. In the Claude apps, upload it under Settings → Capabilities → Skills.
Does wrap-up need an internet connection?
The source says the skill does not use the network. It runs local Python scripts that read your build sheet and goal files.
Can I use wrap-up if I did not use the agent-launcher workflow?
No. The skill expects specific files such as build-sheet.json and goal.json in ./my-agent/, and it will not be useful without them.
What happens if my build sheet is out of date?
The recap is only as accurate as the recorded sheet. Update the build sheet first, because the skill is designed not to claim primitives that were not created.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Wrap-up is a good fit if you launched a Claude Managed Agent through the agent-launcher workflow and want a clear inventory, a shareable overview page and a short upgrade list. If you have not built the agent with that workflow, or your build sheet is out of date, it will not give you a useful result and you should look for another approach.
Download the skill