Skip to content

marketplace-dev

Plugin-author toolkit for Claude Code. Scaffold new plugins and marketplaces, audit any plugin for structural compliance, and maintain existing plugins with confidence.

marketplace-dev has no hard runtime dependency on dev-team — install it on its own to build or maintain plugins.

When to use this

  • Starting a new plugin: the scaffold-plugin skill generates an audit-clean skeleton in one command.
  • Authoring agents or skills: /agent-type-advisor recommends markdown vs. script for a use-case; /agent-create generates a correctly structured agent file.
  • Auditing an existing plugin: the plugin-audit skill produces a structured findings report with zero noise for compliant plugins.
  • Setting up a new marketplace: the scaffold-marketplace skill wires the catalog, release-please config, and at least one plugin slot.

Install

Prerequisites

Required:

  • Claude Code installed and authenticated.
  • jq — used by structural checks.
  • macOS: brew install jq
  • Linux: apt install jq or yum install jq

Optional:

  • git — used for commit-scoped audit snapshots.

Install the plugin

# From the marketplace
claude plugin marketplace add https://github.com/bdfinst/agentic-dev-team
claude plugin install marketplace-dev@bfinster

# From a local clone
claude plugin install --scope project /path/to/agentic-dev-team/plugins/marketplace-dev

Skills & commands

Six skills are user-invocable slash commands; the four scaffolding/audit skills are agent-loaded — Claude dispatches them when the task calls for it rather than you typing a slash command. The Skills catalog is the canonical, generated list with full frontmatter descriptions.

Skill Invocation What it does
agent-type-advisor /agent-type-advisor <prose\|file> Recommend markdown vs. script for a use-case, or audit an existing file
agent-create /agent-create Create an agent file following the official schema and token budgets
agent-skill-authoring /agent-skill-authoring Conventions, anti-patterns, and meta-patterns for authoring agents and skills
agent-add /agent-add Create a new review or team agent (delegates to agent-create)
agent-remove /agent-remove Remove an agent and clean up registry/doc references
add-plugin /add-plugin <name@marketplace> Install a plugin and register it in a project's settings.json
scaffold-plugin agent-loaded Create a new plugin dir with the audit-clean skeleton
scaffold-marketplace agent-loaded Create a marketplace root — catalog, release-please wiring, ≥1 plugin slot
init-plugin-eval agent-loaded Scaffold evals/<name>/{fixtures,expected}/ + grading-contract README
plugin-audit agent-loaded Structural compliance check — agent type, frontmatter, eval coverage, body budgets

Agent

plugin-best-practices-review — read-only, JSON output, structural findings. Checks agent type appropriateness (markdown vs. script), frontmatter compliance, eval-coverage presence, and body line-count budgets. It does not evaluate detection-logic quality — that belongs to the plugin's own agent-eval.

Documentation

Doc Covers
Workflows All commands — scaffolding, agent authoring, maintenance
Skills catalog Full skill/command list with descriptions and options
Agents plugin-best-practices-review agent and dispatch model
Agent-type Decision Rules Markdown vs. script decision matrix (R1–R10)

Conventions enforced

  • Shipping hygiene. Only shipped files live under plugins/<name>/. Eval fixtures and tests live at the repo root (evals/<name>/, tests/).
  • Independent versioning. Each plugin carries its own semver in plugin.json; release-please keeps plugin.json, the release tag, and the catalog entry in lock-step. Do not hand-edit versions.
  • Portability. All shell is #!/usr/bin/env bash, bash 3.2-safe across macOS, Linux, and Git Bash on Windows; every install.sh carries the Git-Bash-on-Windows guard.
  • Audit-clean bar. Scaffolded plugins and migrated skills must pass /plugin-audit with zero findings; plugin-best-practices-review produces zero findings against marketplace-dev itself (dogfood).