Skip to content
Get packs

Stage and Launch Managed Agent

Turns a validated build sheet into API payloads and a resumable launch script you run with your own Anthropic key.

Source Download skill
Repository
alirezarezvani/claude-skills
Stars
29,149
Licence
MIT
Files
5 files

What it does

This skill is the second step in setting up a Claude Managed Agent. It converts a planned build sheet into the exact JSON payloads the Anthropic API expects, writes a launch.sh script that creates the environment, agent, session and first message in order, and runs a validation check that scans for accidentally embedded API keys. The launch script reads your key from your own shell environment, and the skill's tools make no network calls themselves; you run launch.sh yourself.

Great for

  • Developers who have already planned a Claude Managed Agent and want ready-to-run API payloads
  • Founders deploying their own agent who want to use their own Anthropic API key safely
  • Teams that want a repeatable, resumable launch process instead of hand-written curl commands

Skip it if

  • Anyone without an Anthropic API account, since the launch step needs their own key and will incur API usage costs
  • Users who have not first created a validated build sheet, because the skill depends on that input
  • Non-technical users who are uncomfortable running shell scripts in a terminal

Try asking

  • “Launch it: generate the payloads and launch script for my build sheet at ./my-agent/build-sheet.json”
  • “Run the pre-launch validator on my-agent and tell me if anything blocks launch”
  • “Deploy the agent now using my ANTHROPIC_API_KEY from my shell”
Step-by-step guide: how to use Stage and Launch Managed AgentInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 5 original files, the MIT licence and an attribution note.

    stage-launch.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip stage-launch.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip stage-launch.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Run payload_validator.py before launch.sh and fix any FAIL results, especially key_leak findings.
  • Keep MCP tool permissions set to always_ask unless you have reviewed the server.
  • Review launch.sh and the generated payloads before running it, since launch.sh will create billable cloud resources under your account.

What's inside

  • README.md
  • scripts/launch_script_writer.py
  • scripts/payload_generator.py
  • scripts/payload_validator.py
  • SKILL.md

Redistributed unmodified from alirezarezvani/claude-skills under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.