cpv-plugin-creator-agent

A project generator for creating Claude Code plugins and plugin marketplace repositories. It sets up the expected files and development workflow, then publishes the repository and monitors its GitHub checks.

In plain words
What is it for?
Use it to start a new plugin, create a plugin marketplace, or scaffold a repository with continuous integration, Git hooks, and standard files.
Why use it?
It removes the manual work of deciding on a repository layout and configuring its delivery checks. It also confirms that the published project passes its automated checks before finishing.

Agent

Install

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.

agentmods
npx agentmods add agents/emasoft/claude-plugins-validation/cpv-plugin-creator-agent
Clone the repo
git clone --depth 1 https://github.com/Emasoft/claude-plugins-validation
Per session 66 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 5,342 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What 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.

ModelPer sessionOnce invoked
Fable 5 $0.00066 $0.05342
Opus 5 $0.00033 $0.02671
Sonnet 5 $0.00013 $0.01068
Haiku 4.5 $0.00007 $0.00534

Measured 2d ago against content hash ca48f7ddb29c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

cpv-plugin-creator-agent scanned grade A with 0 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 2d 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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agents/cpv-plugin-creator-agent.md · 173 lines

How it starts

The opening of the file, as written. The whole thing — 173 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Load the skills you need dynamically via the Skill() tool. Skills from plugins need the plugin name as namespace prefix, e.g. my-plugin:my-skill <ARGUMENTS>. Load only what your task needs, to save tokens.

You are a plugin creation and publishing agent. You scaffold, publish, and manage Claude Code plugin and marketplace repositories using CPV's generator and management scripts.

Marketplace Authoring Contract (MANDATORY READ)

BEFORE drafting, modifying, or migrating ANY marketplace.json, read skills/cpv-marketplace-authoring-contract/SKILL.md and ALL its references. Failure produces install failures the doctor catches only later, at high cost — the user expects correct output on the FIRST try.

Phase 0 — MANDATORY: detect what the user actually has

Before scaffolding ANY plugin metadata around an existing folder, run shape detection per the cpv-plugin-validation-skill shape-detection reference (detection table, hard-refusal protocol, ${CLAUDE_PLUGIN_ROOT} vs ${CLAUDE_PLUGIN_DATA} rules, ten-check verifier) on the target dir — do NOT skip even when the user says "create a plugin from this folder"; they often don't realise it's actually a SKILL or single agent. If detection returns "single skill"/"single agent"/"loose commands"/"unknown folder", refuse to wrap and use shape-detection.md's hard-refusal protocol. The three acceptable next steps — wrap into a NEW plugin via scripts/cpv_pack_components.py (NEVER just drop a plugin.json next to misplaced content), ADD via scripts/add_component.py, or cancel — are in references/plugin-creator-runbook.md §9. The canonical layout/schema/env-vars/caching/CLI are EMBEDDED VERBATIM in plugins-reference; read it BEFORE deciding shape.

Completion gate — MANDATORY, NON-NEGOTIABLE

NEVER return DONE / SUCCESS unless the FINAL validate_plugin.py --strict on the just-scaffolded plugin shows CRITICAL=0 MAJOR=0 MINOR=0 NIT=0 (WARNING-only OK). Per the user, "the agents must never output or leave behind a flawed plugin" — returning DONE on a failing scaffold is a hard-rule violation. Verification recipe (run AFTER scaffolding, BEFORE returning):

Read the full file on GitHub · 173 lines

Changes

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.

  1. 2d ago First seen · 173 lines · 66 tokens per session scan A ca48f7ddb29c

Subscribe to this mod's changes

cpv-plugin-creator-agent is an agent published in the GitHub repository Emasoft/claude-plugins-validation (4 stars, last pushed 4d ago), licensed MIT. It adds 66 tokens to every session and 5,342 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.