subjective-label
v0.5.1Human-grounded subjective labeling in two sibling sides: Label Building calibrates and freezes one identified human's construct into a signed Label Handoff; Label Scanning qualifies executors on sealed human gold, labels the corpus under that frozen handoff, and closes with an independent probability audit.
By Junjie LuoLicense: MIT1 GitHub starsUpdated last week
Directory evidence
- Runtimes
- Claude Code
- Parsed components
- 6 skill or MCP entries
- Source updated
- Sep 16, 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 subjective-label for Claude Code
claude plugin marketplace add IchenDEV/agent-plugin-mkt
claude plugin marketplace update agent-plugin-marketplace
claude plugin install subjective-label@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/jluo41/ToolsClone the source repository, then follow its setup instructions to add the plugin to a compatible client. The plugin root is plugins/subjective-label/.
Plugin files
├── .claude-plugin/plugin.json├── skills/label-building/SKILL.md├── skills/label-building-workflow/SKILL.md├── skills/label-scanning/SKILL.md├── skills/label-scanning-workflow/SKILL.md├── skills/subjective-label/SKILL.md└── skills/subjective-label-workflow/SKILL.md
Included Skills6
The Building-side door of the subjective-label family, and its LAW: who is the semantic authority, what each of Contract, Round and Freeze may and may not create, which decisions are human gates, which verbs exist, and what is forbidden. It ends at a signed Label Handoff and never evaluates executors or scans the corpus. The order of steps lives in label-building-workflow. Use for new labeling jobs, calibration rounds, human annotation sessions, boundary discovery, guideline revision, stopping decisions, freeze, or /label-building.
The ORDER machine of the Building side of the subjective-label family: drives P0 Contract, the P1 Round loop (PREPARE, JUDGE, LEARN, CLOSE), and P2 Freeze as independently closable Labeling operations on disk, owns item-level resume and Run receipts, and hands the crossing back to subjective-label-workflow. It owns no law: authority, human gates and forbidden acts live in label-building. Use when running or resuming a calibration round, opening a round card, resuming a Session, closing a checkpoint, or /label-building-workflow.
The Scanning-side door of the subjective-label family, and its LAW: what a frozen Label Handoff permits, what Test, Scan and Audit may and may not create, which decisions are human gates, which verbs exist, and what is forbidden. It begins from one signed handoff and ends at audited D*. The order of steps lives in label-scanning-workflow. Use for sealed evaluation, model qualification, production labeling, risk queues, corpus scanning, spot checks, repair, final audit, or /label-scanning.
The ORDER machine of the Scanning side of the subjective-label family: drives P3 Test (GOLD locked before SCORE), P4 Scan (manifest, preflight, shards, risk queue, reconcile), and P5 Audit (design, blind sample, findings, route, repair loop) as independently closable Labeling operations on disk, owns Run-level resume and receipts, and hands the crossing back to subjective-label-workflow. It owns no law: boundary, human gates and forbidden acts live in label-scanning. Use when running or resuming an evaluation, a production scan, a risk queue, an audit or a repair, or /label-scanning-workflow.
The one user-facing door for a human-grounded subjective-label family with two sides: Label Building teaches and freezes what one identified human means by a vague trait; Label Scanning qualifies executors against a sealed human-gold test, scans the corpus under the frozen meaning, and audits the completed labels. Use for starting, resuming, routing, or checking a subjective annotation job; defining H/L/N and boundary regions; calibration rounds; guideline freeze; executor evaluation; corpus scanning; final audit; or /subjective-label.
The crossing machine of the subjective-label family: declares the six phase names P0-P5 and the seven gates G0-G6, tests Label Handoff validity, decides which side is runnable, routes invalidation between sides, and dispatches one bounded action to label-building-workflow or label-scanning-workflow. It owns no step order, no semantic rule, no item judgment, no executor prediction, and no audit verdict. It publishes the family's Phase × Run Map. Use when asking where a labeling job is, whether it may cross, why it is blocked, or /subjective-label-workflow.
Plugin manifests1
{
"name": "subjective-label",
"version": "0.5.1",
"description": "Human-grounded subjective labeling in two sibling sides: Label Building calibrates and freezes one identified human's construct into a signed Label Handoff; Label Scanning qualifies executors on sealed human gold, labels the corpus under that frozen handoff, and closes with an independent probability audit.",
"author": {
"name": "Junjie Luo",
"url": "https://github.com/jluo41"
},
"license": "MIT",
"keywords": [
"annotation",
"subjective",
"human-grounded",
"labeling",
"sealed-test",
"human-in-the-loop",
"provenance",
"audit"
]
}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.
[subjective-label on Agent Plugins Marketplace](https://pluginsmp.com/plugins/subjective-label)