- Home
- Skills
- Agents & Orchestration
- Claude Plugin Packager Pro: Package, Validate, Ship
More screenshots
Works with the AI Agents you already use
Claude Plugin Packager Pro: Package, Validate, Ship
Turning a folder of skills into something other people can install is fiddly: manifest fields, hook paths that break on other machines, executable bits, secrets that slip in.
$9
Claude Plugin Packager Pro: Package, Validate, Ship
Example session with this skill installed
Package my .claude skills, commands and hooks into a Claude Code plugin I can publish on GitHub, and check it for problems first.
- Read your context and instructions
- Compiled the claude plugin packager
Packaged pk-test from .claude/ into ./marketplace and validated it. First run found 7 errors; after fixes it's clean and bumped to 1.1.0.
Package
$ python3 scripts/package.py --name pk-test --author Igna --out ./marketplace --description "Test plugin"
Packaged pk-test into ./marketplace: skills=1, commands=1, agents=0, hooks=4
Validate (first run)
ERROR marketplace/pk-test/agents/a.md: agent needs name
ERROR marketplace/pk-test/hooks/block-dangerous.sh: hook script not executable (chmod +x)
ERROR marketplace/pk-test/.env: env file must not ship in a plugin
WARN marketplace/pk-test/agents/a.md: no tools list: agent inherits ALL tools
1 plugin(s), 7 error(s), 1 warning(s)
| Problem | Fix |
|---|---|
Agent without name | Added frontmatter name and a tools list |
| Hook scripts not executable | chmod +x hooks/*.sh |
.env inside the plugin | Removed |
Validate (after fixes) and release
1 plugin(s), 0 error(s), 0 warning(s)
$ python3 scripts/bump.py ./marketplace/pk-test minor "Added hooks"
pk-test -> 1.1.0
Next steps
/plugin marketplace add ./marketplaceand/plugin install pk-test@pk-test-marketplaceto test locally.- Push to GitHub; the included workflow validates every PR and tags new versions.
Connects securely to your tools. The creator never sees your data.
About this skill
Turning a folder of skills into something other people can install is fiddly: manifest fields, hook paths that break on other machines, executable bits, secrets that slip in. This skill does the packaging and checks it before you publish.
What you get
| Script | What it does |
|---|---|
package.py | Copies chosen skills, commands, agents, hooks and .mcp.json from your .claude/ folder into a plugin, rewrites hook paths to ${CLAUDE_PLUGIN_ROOT}, writes plugin.json, marketplace.json, README, CHANGELOG and a GitHub Actions workflow. |
validate.py | Checks manifests (kebab-case name, semver), skill frontmatter (name matches folder, description length), agent tools, hook JSON, missing or non-executable hook scripts, project/absolute paths, .env files, absolute user paths and common secret patterns (Anthropic, OpenAI, AWS, GitHub, Slack, Google keys, private keys). Exits non-zero on errors. |
bump.py | Bumps major/minor/patch in plugin.json and the marketplace entry and adds a dated CHANGELOG entry. |
template/ | Starter marketplace with an example skill, command, agent, hook and release workflow that tags versions. |
Pure Python 3.9+, no dependencies.
How it works
- Claude lists what's in your
.claude/folder and you pick what ships. package.py --name my-plugin --author "You" --out ./marketplacevalidate.py ./marketplace: fix every ERROR.- Test locally:
/plugin marketplace add ./marketplace, then/plugin install my-plugin@<marketplace>. bump.pyfor each release; push to GitHub so users can add<user>/<repo>.
Limits
The secret scan catches common key formats, not every possible secret: still review before publishing. If Claude Code changes the plugin manifest format, check the current docs.
Requires Python 3.9+. Cannot automatically resolve external library dependencies for packaged scripts. Validation requires manual fix for detected secrets or absolute paths.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 4 days ago
Requires