Agent Plugins Marketplace
← All plugins

goalkeeper

v0.7.0

Durable goal-following for Claude Code: contracts with definition-of-done, subagent judge gates, executor-subagent chain execution, and adaptive missions — with all state mechanics executed by a tested CLI and invariants enforced by hooks. Inspired by OpenAI Codex /goal and the Ralph loop pattern.

Claude Code8 Skills

By Youssef HemimyLicense: MIT13 GitHub starsUpdated 1 hour ago

Directory evidence

Runtimes
Claude Code
Parsed components
8 skill or MCP entries
Source updated
Aug 29, 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 plugin

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

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

Plugin files

goalkeeper/
├── .claude-plugin/plugin.json
├── skills/goal/SKILL.md
├── skills/goal-chain/SKILL.md
├── skills/goal-clear/SKILL.md
├── skills/goal-judge/SKILL.md
├── skills/goal-pause/SKILL.md
├── skills/goal-prep/SKILL.md
├── skills/goal-resume/SKILL.md
└── skills/goal-supervisor/SKILL.md

Included Skills8

goalskills/goal/SKILL.md

Set or check status of a durable goalkeeper goal. Use this skill when the user invokes /goal "<objective>" to start a new contract-driven goal, or /goal with no arguments to see status of the currently active goal. Goalkeeper goals run autonomously across many turns with checkpoint validation and judge-gated completion.

goal-chainskills/goal-chain/SKILL.md

Run a linear sequence of goalkeeper goals where the judge gates progression between them. Use when the user invokes /goal-chain "<file>" to start a chain. Chain bookkeeping (cursor, approvals, activation) is executed by the gk CLI.

goal-clearskills/goal-clear/SKILL.md

Stop and archive the currently active goalkeeper goal. Use when the user invokes /goal-clear to abandon or finalize a goal. Files are moved to .claude/goals/_archive/, never deleted.

goal-judgeskills/goal-judge/SKILL.md

The gate. Reviews the active goalkeeper goal against its definition-of-done and either approves (advance / mark done) or rejects (with a structured fix-list). Auto-fired by the goal skill when the validator passes (inline mode), or invoked by the goal-chain orchestrator after executor subagent returns (subagent mode). Can also be invoked on demand via /goal-judge for advisory review.

goal-pauseskills/goal-pause/SKILL.md

Pause the currently active goalkeeper goal without losing state. Use when the user invokes /goal-pause. The goal can later be resumed with /goal-resume.

goal-prepskills/goal-prep/SKILL.md

Interactively draft a goalkeeper contract before executing. Use this skill when the user invokes /goal-prep "<rough idea>", or when /goal is called for a slug that has no contract yet. Produces a well-formed contract.md with explicit objective, definition-of-done, validator command, and non-goals.

goal-resumeskills/goal-resume/SKILL.md

Resume a paused or needs_human goalkeeper goal. Use when the user invokes /goal-resume after they've manually unblocked the goal (e.g. fixed a problem the judge flagged).

goal-supervisorskills/goal-supervisor/SKILL.md

The mission-level supervisor. One level above goals. Reads the user's mission charter and the most-recently-completed goal's artifacts, then decides whether to PROCEED (draft the next goal's contract for user review — never auto-activate), declare the mission DONE, or ESCALATE to the user. Use this skill when the user invokes /goal-supervisor after a goal completes, or when running a multi-goal arc where each next goal's shape is informed by the prior goal's actual output.

Plugin manifests1

.claude-plugin/plugin.json
{
  "name": "goalkeeper",
  "version": "0.7.0",
  "description": "Durable goal-following for Claude Code: contracts with definition-of-done, subagent judge gates, executor-subagent chain execution, and adaptive missions — with all state mechanics executed by a tested CLI and invariants enforced by hooks. Inspired by OpenAI Codex /goal and the Ralph loop pattern.",
  "author": {
    "name": "Youssef Hemimy",
    "email": "[email protected]",
    "url": "https://youssefhemimy.dev"
  },
  "homepage": "https://github.com/bonfire-systems/goalkeeper",
  "repository": "https://github.com/bonfire-systems/goalkeeper",
  "license": "MIT",
  "keywords": [
    "claude-code",
    "claude-code-plugin",
    "agents",
    "autonomous",
    "goals",
    "automation"
  ]
}

For maintainers

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

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