Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/phazurlabs/install-labs/claude-code-plugin-packagingnpx skills add phazurlabs/install-labs --skill claude-code-plugin-packaginggit clone --depth 1 https://github.com/phazurlabs/install-labsWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00097 | $0.04690 |
| Opus 5 | $0.00048 | $0.02345 |
| Sonnet 5 | $0.00019 | $0.00938 |
| Haiku 4.5 | $0.00010 | $0.00469 |
Grade F, and why
claude-code-plugin-packaging scanned grade F with 4 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
curl -fsSL https://raw.githubusercontent.com/you/my-plugin/main/install.sh | bash Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
cat > ~/.claude/plugins/$PLUGIN/.claude-plugin/plugin.json << 'EOF' Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
rm -rf ~/.claude/plugins/my-plugin Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -fsSL https://raw.githubusercontent.com/you/my-plugin/main/install.sh | bash How it starts
The opening of the file, as written. The whole thing — 600 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Claude Code Plugin Packaging
Complete reference for building, structuring, distributing, and maintaining Claude Code plugins — the native extension format for Claude Code.
What Is a Claude Code Plugin?
A Claude Code plugin is a directory installed at ~/.claude/plugins/<name>/ that extends Claude Code with three capabilities:
| Component | Purpose | Invocation |
|---|---|---|
| Skills | Domain knowledge auto-injected into conversations | Automatic — keyword match in the skill's description field triggers inclusion |
| Commands | Structured protocols the user triggers explicitly | Manual — user types /command-name in the conversation |
| MCP Servers | Tool servers providing actions Claude can call | Automatic — declared in plugin.json, started when the plugin loads |
A plugin can contain any combination of these. A pure-knowledge plugin might have only skills. A workflow plugin might have only commands. A tool plugin might bundle an MCP server with contextual skills that teach Claude how to use it effectively.
Key mental model: Skills are what Claude knows. Commands are what Claude does on request. MCP servers are what Claude can interact with.
When to Build a Claude Code Plugin
Build a plugin when you want to:
- Bundle domain expertise — Give Claude deep knowledge in a specific area (medical coding, legal research, design systems) that activates automatically when relevant
- Create repeatable workflows — Define multi-step protocols (code review checklists, deployment procedures, audit processes) invokable via slash commands
- Package MCP tools with context — An MCP server alone gives Claude tools; a plugin wraps those tools with skills that teach Claude when and how to use them
- Distribute to a team — Share a standardized Claude Code configuration across developers, designers, or analysts
- Sell on a marketplace — Package expertise as a product others install with one command
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago First seen · 600 lines · 97 tokens per session scan F 084177e858f8
claude-code-plugin-packaging is a skill published in the GitHub repository phazurlabs/install-labs (3 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 97 tokens to every session and 4,690 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it F with 4 findings (downloads and executes remote code, reads agent configuration directories, recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
promptscript
PromptScript language expert for reading, writing, modifying, and troubleshooting .prs files. Use when working with PromptScript syntax, creating or editing .prs files, adding blocks like @identity, @standards, @restrictions, @shortcuts, @skills, or @agents, configuring promptscript.yaml, resolving compilation errors…
tdd-workflow
Test-driven development workflow.
expert
Base expert skill.
alpha
Alpha skill.
local-ci-dev
Run local CI via the in-tree dev build of local-ci (pnpm local-ci-dev) to verify changes to this repo before completing work. Runs pnpm local-ci-dev run --all in the background, watches the log for step failures, and retries failed runners after fixes. Use before reporting work as complete, or whenever the user asks…
local-ci
Run GitHub Actions workflows locally with pause-on-failure for AI-agent-driven CI iteration.