Tier Label — Development skill for Claude Code
人工标注 tier-guard 的判定：R-IRREVERSIBLE 命中是否误报（fp/tp）、T1 产出是否被打回（reject/accept）.
How to install Tier Label
Installs to ~/.claude/skills/haigeer-labs-tier-guard-tier-label/SKILL.md
mkdir -p ~/.claude/skills/haigeer-labs-tier-guard-tier-label && curl -fsSL https://raw.githubusercontent.com/haigeer-labs/tier-guard/HEAD/commands/tier-label.md -o ~/.claude/skills/haigeer-labs-tier-guard-tier-label/SKILL.md Restart Claude Code, or start a new session, for it to be picked up.
What Tier Label does
description: 人工标注 tier-guard 的判定：R-IRREVERSIBLE 命中是否误报（fp/tp）、T1 产出是否被打回（reject/accept）
argument-hint: " "
allowed-tools: Bash
参数：`$ARGUMENTS`
**只接受用户在本轮亲口给出的判定。** 参数不全、或判定是你推断出来的，就停下来问用户， 不要自己补 —— 误报率和打回率决定能不能切 auto，模型替人标注等于自己给自己的守卫放行。
参数齐全时，把 `
python3 "${CLAUDE_PLUGIN_ROOT}/hooks/tier_label.py" "" "<判定>" --data "${CLAUDE_PLUGIN_DATA}" 把输出原样转告用户。报错（id 不存在、类型对不上）就照实说，不要换个 id 重试。
Alternatives in Development
- Label Issue 1.3k ★
- Trust Tiers — Trust Tier Badges 325 ★
- Skill Tiers — Skill Tier Taxonomy 311 ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
Proposals
List open scout-proposed features; lets you triage (accept → develop, reject → suppress)
Tier Doctor
Claude Code 中只读诊断 tier-guard 的数据目录、日志与 Codex hook 可观察性
Tier Report
tier-guard 路由摘要：v2 请求/选择/派发/实际观测，以及历史 v1 记录
Kelly Advisor
A survival-first Kelly Criterion decision skill for Claude. Sizes your bets, defaults to fractional Kelly, and
Guillotines
Surface all [GT] (guillotine) tagged items across the vault — hard-deadline tasks that foreclose an option or
Verify Effect
Verify the effect of an irreversible action — probe the resulting state, not the action's success report.
Related Agents
Code Preview.Nvim
A Neovim plugin that shows a live diff preview of AI coding agent's file edits before you accept or reject the
Scope Guard (Shared Prompt Block For Tier 2 Agents)
This file is not a standalone agent. It contains the shared scope enforcement prompt text that Tier 2 (executi
