Agent Plugins Marketplace
← All plugins

feature

v0.9.24

Ship a feature from a written plan: the plan skill spawns the planner and gates its decisions, open questions and risks with you interactively; the ship skill runs the codex and UX lanes to a merged, verified, deployed PR; the planner, ux-implementer, ux-autofixer and ux-verifier agents staff the lanes; the handoff skill records a mid-phase stop; the retro skill ranks a finished session's biggest wastes by token cost and routes fixes back to ground, the planner and ship; hooks keep subagents and the main loop from blocking on each other.

Claude Code5 Skills

By blockchainianLicense: MIT0 GitHub starsUpdated 1 hour ago

Directory evidence

Runtimes
Claude Code
Parsed components
5 skill or MCP entries
Source updated
Sep 27, 2026
Manifest status
Canonical path parsed

The directory validates manifest shape and source location. It does not execute the plugin or provide a security endorsement. Review the indexing methodology →

Install feature for Claude Code

Installs for the current user
claude plugin marketplace add IchenDEV/agent-plugin-mkt
claude plugin marketplace update agent-plugin-marketplace
claude plugin install feature@agent-plugin-marketplace

Paste and run these commands in a terminal with Claude Code. They add and refresh the PluginsMP catalog, then install this plugin.

The installer fetches third-party code from the source repository shown on this page. This directory validates manifest structure and source location, but does not perform a security audit; review the manifest, components, and source before installing.

Get the source manually
git clone https://github.com/blockchainian/claude

Clone the source repository, then follow its setup instructions to add the plugin to a compatible client. The plugin root is plugins/feature/.

Plugin files

plugins/feature/
├── .claude-plugin/plugin.json
├── skills/ground/SKILL.md
├── skills/handoff/SKILL.md
├── skills/plan/SKILL.md
├── skills/retro/SKILL.md
└── skills/ship/SKILL.md

Included Skills5

groundskills/ground/SKILL.md

Pin repo facts BEFORE a planning turn, so the planner transcribes checked facts instead of inferring them. Use when about to plan or design any change against an existing codebase — "plan X", "design Y", "write a spec for Z" — and especially before handing a spec to /codex:implement, whose gates test whether code passes, never whether a premise was true. Produces a grounding doc and stops before design. NOT for planning itself, and not for changes small enough to hold in one turn.

handoffskills/handoff/SKILL.md

Write a mid-phase handoff when a session must stop before its phase's output exists — quota exhausted, end of day, the context safety rail in the user's CLAUDE.md, or the 55-minute idle wake-up firing. Use for "/feature:handoff", "write a handoff", or when the idle wake-up prompt says to. NOT at a phase end: a phase's output (problem.md, plan.md, the PR, verdict JSONs, the report) is its own record and needs no handoff.

planskills/plan/SKILL.md

Run the feature pipeline's planning phase: spawn the planner on a grounded problem.md, then interactively approve its decisions, open questions and risks before ship. Use after /feature:ground and before /feature:ship — "plan X", "run the planner", "review the plan's decisions". This is the feature pipeline's plan step, NOT Claude Code's built-in /plan (plan mode). NOT for grounding, not for shipping, and not for a change small enough to hold in one turn.

retroskills/retro/SKILL.md

Find the biggest wastes of effort in a finished orchestration session and route each fix back into the workflow it came from. Run in a SEPARATE session, passing the finished session's name: "/feature:retro <session-name>". Reads that session's transcript as evidence, ranks wastes by real token cost, and proposes fixes to ground, the planner and ship. NOT in the session being analysed (a session grading itself inherits the blind spot that caused the waste), and not for a one-turn task with nothing to rank.

shipskills/ship/SKILL.md

Run a written plan through the codex and UX lanes to a shipped feature — launch the workstreams, write and run the probes, verify staging, triage the local review, run the fix lanes, decide production. Use when a `plan.md` exists and has passed the path checker: "/feature:ship <plan.md>", "run this plan", "ship this plan". NOT for planning, and not for a change small enough to do in one turn — there the launch overhead is the whole cost.

Plugin manifests1

plugins/feature/.claude-plugin/plugin.json
{
  "name": "feature",
  "version": "0.9.24",
  "description": "Ship a feature from a written plan: the plan skill spawns the planner and gates its decisions, open questions and risks with you interactively; the ship skill runs the codex and UX lanes to a merged, verified, deployed PR; the planner, ux-implementer, ux-autofixer and ux-verifier agents staff the lanes; the handoff skill records a mid-phase stop; the retro skill ranks a finished session's biggest wastes by token cost and routes fixes back to ground, the planner and ship; hooks keep subagents and the main loop from blocking on each other.",
  "author": {
    "name": "blockchainian"
  },
  "license": "MIT",
  "homepage": "https://github.com/blockchainian/claude",
  "repository": "https://github.com/blockchainian/claude",
  "keywords": [
    "planning",
    "orchestration",
    "codex",
    "ux",
    "verification",
    "handoff"
  ]
}

If you maintain this plugin, link to this source-backed listing from your README so users can review its manifest and indexed components.

[feature on Agent Plugins Marketplace](https://pluginsmp.com/plugins/feature)