fankeel
v0.70.0Development discipline for long-running projects: a seven-stage pipeline restated on every prompt and after every answer with the shape its report takes, a per-session task registry, capped task memory, a brief for every subagent, and warnings — optionally blocks — when two live sessions are editing the same files.
By FanFantom94520 GitHub starsUpdated 6 days ago
Directory evidence
- Runtimes
- Claude Code
- Parsed components
- 11 skill or MCP entries
- Source updated
- Sep 18, 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 fankeel for Claude Code
claude plugin marketplace add IchenDEV/agent-plugin-mkt
claude plugin marketplace update agent-plugin-marketplace
claude plugin install fankeel@agent-plugin-marketplacePaste 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/FanFantom9452/FanKeelClone 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
├── .claude-plugin/plugin.json├── skills/fankeel/SKILL.md├── skills/fankeel-ask/SKILL.md├── skills/fankeel-audit/SKILL.md├── skills/fankeel-build/SKILL.md├── skills/fankeel-design/SKILL.md├── skills/fankeel-explain/SKILL.md├── skills/fankeel-land/SKILL.md├── skills/fankeel-plan/SKILL.md├── skills/fankeel-station/SKILL.md├── skills/fankeel-survey/SKILL.md└── skills/fankeel-verify/SKILL.md
Included Skills11
Task registry and development discipline for long-running projects. Use for /fankeel, starting or pausing a task, asking what this or another session is working on, moving to the next stage, or the station — "show all sessions", "which sessions are still open", "clean up old sessions", "監控站". Runs a task through a route it picks from survey, design, plan, build, verify, audit and land, and warns — optionally blocks — when another live session shares your files.
Interrupt the stage you are in and put one question to a stronger model — a one-shot judge whose answer is filed verbatim and then acted on. Use for /fankeel-ask, "問判官", "叫 fable", "這題我們解不開", or when a mid-stage question is worth a model above the one running the session.
Audit documentation against the code it describes — dead references, pages that stopped being true, two pages describing one thing, plans whose work has landed, orphans. Use for /fankeel-audit, "check the docs", "what is out of date", "文件審查", before a release, or when two documents disagree.
The build stage — run a plan's tasks, or a design's file table where there is no plan, in a loop that does not stop to ask, keeping its place in a ledger and reviewing each task as it lands. Use for the build stage of a fankeel task, implementing an approved plan, resuming build work after a compaction, or when a task loop needs a ledger.
The design stage — one approach with its trade-offs, a success criterion that can fail, and a check against the project map before anything is built. Use for the design stage of a fankeel task, choosing between approaches, writing a spec, or when an approach needs approving before implementation.
Say it so it is understood on first reading — the one sentence first, unknowns left unknown, contrast only where evidence has one, a report as a path, a status sync in six fields, and a check before sending. Use for a presentation, a report, a project status sync or init, sorting out a line of thought, 簡報, 報告, 進度同步, 整理思路, or when a session has drifted, repeated itself, or may have misread the project.
The land stage — a green suite, the documents closed, the map rewritten, and the integration decision left to the user. Use for the land stage of a fankeel task, finishing a development branch, deciding between merge and PR, or cleaning up a worktree when work is complete.
The plan stage — decompose an approved design into tasks someone with no context could execute, with constraints generated from the project rather than remembered. Use for the plan stage of a fankeel task, writing an implementation plan, or breaking a spec into tasks before any code is written.
Run the station as a live server — every fankeel session on this machine on one page, served rather than read as the static file a prompt already writes. Use for /fankeel-station, "開站", or when the station has to be a server rather than a file. Reading that file, and the station's other phrases like "監控站", stay with the fankeel skill.
The survey stage — read the project's own map before reading its code, classify the work, and report what is already here. Use for the survey stage of a fankeel task, "what is already here", starting work in an unfamiliar repository, or when a task needs classifying as spike, bounded or architectural.
The verify stage — evidence before claims, requirements checked line by line, and the documents this change just made false. Use for the verify stage of a fankeel task, before claiming work is complete or passing, before a commit or PR, or when checking whether a change broke the documentation describing it.
Plugin manifests1
{
"name": "fankeel",
"version": "0.70.0",
"description": "Development discipline for long-running projects: a seven-stage pipeline restated on every prompt and after every answer with the shape its report takes, a per-session task registry, capped task memory, a brief for every subagent, and warnings — optionally blocks — when two live sessions are editing the same files.",
"author": {
"name": "FanFantom9452",
"url": "https://github.com/FanFantom9452"
},
"skills": "./skills/",
"agents": [
"./agents/fankeel-reader.md",
"./agents/fankeel-judge.md",
"./agents/fankeel-reviewer.md",
"./agents/fankeel-verifier.md",
"./agents/fankeel-fixer.md"
],
"hooks": {
"SessionStart": [
{
"matcher": "clear|fork",
"hooks": [
{
"type": "command",
"command": "node \"${CLAUDE_PLUGIN_ROOT}/hooks/carry.js\"",
"timeout": 5,
"statusMessage": "Looking for the task the clear left behind..."
}
]
}
],
"UserPromptSubmit": [
{
"hooks": [
{
"type": "command",
"command": "node \"${CLAUDE_PLUGIN_ROOT}/hooks/inject.js\"",
"timeout": 5,
"statusMessage": "Reading the fankeel registry..."
}
]
}
],
"PostToolUse": [
{
"matcher": "AskUserQuestion",
"hooks": [
{
"type": "command",
"command": "node \"${CLAUDE_PLUGIN_ROOT}/hooks/resume.js\"",
"timeout": 5,
"statusMessage": "Restating the fankeel stage..."
}
]
},
{
"matcher": "Edit|Write|NotebookEdit",
"hooks": [
{
"type": "command",
"command": "node \"${CLAUDE_PLUGIN_ROOT}/hooks/touch.js\"",
"timeout": 5,
"statusMessage": "Noting where the work went..."
}
]
},
{
"hooks": [
{
"type": "command",
"command": "node \"${CLAUDE_PLUGIN_ROOT}/hooks/size.js\"",
"timeout": 5,
"statusMessage": "Checking how much that just added to context..."
}
]
}
],
"SubagentStart": [
{
"hooks": [
{
"type": "command",
"command": "node \"${CLAUDE_PLUGIN_ROOT}/hooks/brief.js\"",
"timeout": 5,
"statusMessage": "Briefing the subagent..."
}
]
}
],
"PreToolUse": [
{
"matcher": "Edit|Write|NotebookEdit",
"hooks": [
{
"type": "command",
"command": "node \"${CLAUDE_PLUGIN_ROOT}/hooks/guard.js\"",
"timeout": 5,
"statusMessage": "Checking the fankeel scope guard..."
}
]
},
{
"matcher": "AskUserQuestion",
"hooks": [
{
"type": "command",
"command": "node \"${CLAUDE_PLUGIN_ROOT}/hooks/gate.js\"",
"timeout": 5,
"statusMessage": "Noting when the gate opened..."
}
]
},
{
"matcher": "Bash|PowerShell",
"hooks": [
{
"type": "command",
"command": "node \"${CLAUDE_PLUGIN_ROOT}/hooks/guard.js\"",
"timeout": 5,
"statusMessage": "Checking whether a read-only agent is about to write..."
}
]
}
],
"SessionEnd": [
{
"hooks": [
{
"type": "command",
"command": "node \"${CLAUDE_PLUGIN_ROOT}/hooks/leave.js\"",
"timeout": 60,
"async": true,
"statusMessage": "Recording how the session ended..."
}
]
}
]
}
}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.
[fankeel on Agent Plugins Marketplace](https://pluginsmp.com/plugins/fankeel)