judge-field-audit
判官字段审计与提交级裁决：交付物字段分数值类 vs 字符串枚举类（同权重受检），枚举字段逐 label 码表核对；判官分解存在多等价解或判官口径分歧时，用破坏性提交级差分（故意弄坏单字段看 harbor…
Skill by Hab1nA
Chalk Labs
Need help deploying judge-field-audit?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/hab1na-judge-field-audit
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: security
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/judge-field-audit
What's inside
- Skill name
judge-field-audit- Description
- 判官字段审计与提交级裁决：交付物字段分数值类 vs 字符串枚举类（同权重受检），枚举字段逐 label 码表核对；判官分解存在多等价解或判官口径分歧时，用破坏性提交级差分（故意弄坏单字段看 harbor 变不变）裁决。差分必须覆盖全部字段类型。触发词：'字段审计'、'枚举字段'、'多等价解'、'decomposition ambiguity'、'判官口径分歧'、'破坏性差分'、'弄坏字段'。
- License
- unknown
- Instructions
- 61 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
The license isn't open source, so the instructions aren't shown here. Read SKILL.md on GitHub
More from Hab1nA/ASCodex
- 0.519score
dft-convergence
Hab1nA
Systematic convergence testing for density functional theory calculations: sweep plane-wave cutoff, k-point grid, smearing width, and supercell size to ensure…
SkillBFreshness: BClaude Code +9 - 0.519score
checkpoint
Hab1nA
Save current task progress to structured checkpoint file.
SkillBFreshness: BClaude Code +9 - 0.519score
distill
Hab1nA
Extract recurring patterns from trace files and distill them into reusable knowledge graph entries: scan traces for errors, surprises, and parameter insights…
SkillBFreshness: BClaude Code +9 - 0.519score
grade-reproduction
Hab1nA
Grade a completed paper reproduction using the Reproduction Quality Score (RQS, 0-110 pts): evaluate figure coverage, model fidelity, quantitative and…
SkillBFreshness: BClaude Code +9 - 0.516score
platform-scorecard-analyze
friday-team
识别 Bohrium Playground 判分器类型，归一化原始分、有效分、判罚、归属与榜单范围，并在授权预算内设计单字段 A/B。高分未满时先定位丢分分量，禁止盲试。触发词：'判分器类型'、'harbor 档位'、'单字段 A/B'、'怎么提分'、'分量级分解'、'哪个分量掉分'。
SkillBFreshness: BClaude Code +9 - 0.516score
closure-evidence-standard
Hab1nA
封板证据标准：solver 建议收关/封板前必须过封板三问（场上有人在你上面？多独立证伪还是只有你？本地/历史曾达更高值？）+ 真天花板核对清单（曾达记录考古、采样 regime 共享遗漏、判分器两轨确认、练习轮额度止损）。本地曾达更高值 =…
SkillBFreshness: BClaude Code +9 - 0.516score
differential-scoring
Hab1nA
差分判分定位：当总分高而未满（≥90% 但不到 100）、且本地各分量都'对'却不知哪块丢分时，用差分实验定位丢分分量/字段，再定点清除。核心：先定位后修复，禁止盲试。触发词：'差分判分'、'哪个分量掉分'、'定位丢分'、'高分未满'、'4% 缺口'、'差一点满分'。
SkillBFreshness: BClaude Code +9 - 0.516score
oracle-probe
friday-team
判官只读诊断：不改内容，从已授权的 worker/判分器侧读出分量级评分、运行日志、逐项校验结果与 scorecard 明细，用于差分判分定位；配额语义以当前服务契约核实。与提交路径分离。触发词：'判官明细'、'oracle probe'、'读判官日志'、'分量级评分'、'scorecard 明细'、'判分器侧'。
SkillBFreshness: BClaude Code +9 - 0.516score
playground-solve-optimal
Hab1nA
Bohrium Playground 得分战术总纲：三类题（确定性 verifier / LLM judge / 预测成像）的最优解法范式与开题到满分的全流程 checklist。触发词：'得分战术'、'最优解法'、'怎么拿满分'、'解题流程'、'playground 攻略'。
SkillBFreshness: BClaude Code +9 - 0.516score
mp-r-family-solve
Hab1nA
mp-r 同家族题（split coannihilation / UV portal 等）的字段级专项规则：裸值 vs 角平均陷阱、保留名三分类、κ_D 符号保留、因子归属、cross_target 继承链。来自 split 15→81.55 与 uv 34→42…
SkillBFreshness: BClaude Code +9 - 0.515score
generate-grader
Hab1nA
Read a hackathon topic markdown file, extract scoring criteria and reference values, and generate a programmatic grader function for…
SkillBFreshness: BClaude Code +9 - 0.513score
competition-coordinate
Hab1nA
总负责人多代理协调作战手册：六类角色分工、OODA 攻坚环、飞书群情报员 hook 驱动三级研判、身份配额中央注册、决策日志、跨代理信息纪律。触发词：'多代理'、'协调'、'总负责人'、'作战'、'派活'、'换人'、'飞书群监控'。
SkillBFreshness: BClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 0
- Last commit (this folder)
- Aug 29, 2026 (2 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- License unknown
- Repository
- github.com/Hab1nA/ASCodex
- Website
- play.bohrium.com/
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from Hab1nA
- 0.519score
checkpoint
Hab1nA
Save current task progress to structured checkpoint file.
SkillBFreshness: BClaude Code +9 - 0.519score
dft-convergence
Hab1nA
Systematic convergence testing for density functional theory calculations: sweep plane-wave cutoff, k-point grid, smearing width, and supercell size to ensure…
SkillBFreshness: BClaude Code +9 - 0.519score
distill
Hab1nA
Extract recurring patterns from trace files and distill them into reusable knowledge graph entries: scan traces for errors, surprises, and parameter insights…
SkillBFreshness: BClaude Code +9 - 0.519score
grade-reproduction
Hab1nA
Grade a completed paper reproduction using the Reproduction Quality Score (RQS, 0-110 pts): evaluate figure coverage, model fidelity, quantitative and…
SkillBFreshness: BClaude Code +9 - 0.516score
closure-evidence-standard
Hab1nA
封板证据标准：solver 建议收关/封板前必须过封板三问（场上有人在你上面？多独立证伪还是只有你？本地/历史曾达更高值？）+ 真天花板核对清单（曾达记录考古、采样 regime 共享遗漏、判分器两轨确认、练习轮额度止损）。本地曾达更高值 =…
SkillBFreshness: BClaude Code +9 - 0.516score
differential-scoring
Hab1nA
差分判分定位：当总分高而未满（≥90% 但不到 100）、且本地各分量都'对'却不知哪块丢分时，用差分实验定位丢分分量/字段，再定点清除。核心：先定位后修复，禁止盲试。触发词：'差分判分'、'哪个分量掉分'、'定位丢分'、'高分未满'、'4% 缺口'、'差一点满分'。
SkillBFreshness: BClaude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get judge-field-audit running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
judge-field-audit skill. Agent Mart. https://agentmart.dev/skills/hab1na-judge-field-audit (last updated 2026-10-24)