Agent skill

Judge

Evaluate JSON state with Noul, Choice or Score questions for ticket triage, classification and graded judgments through the selected judge provider (TypeSafe JEV by default), or cancel an active…

Apache-2.0Auto-check passedSales & Support

Install Judge

skills CLI

$ npx skills add iii-hq/workers --skill judge -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI

$ gh skill install iii-hq/workers judge --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy

$ git clone --depth 1 https://github.com/iii-hq/workers.git skills-src && mkdir -p .claude/skills && cp -r skills-src/judge/skills .claude/skills/judge && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Paste into Claude Code

Install the "judge" agent skill from https://github.com/iii-hq/workers/tree/main/judge/skills into .claude/skills/judge/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "judge", then confirm the skill loads.

Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

$skill-installer

$skill-installer install https://github.com/iii-hq/workers/tree/main/judge/skills

Type this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.

skills CLI

$ npx skills add iii-hq/workers --skill judge -a codex

Project install goes to .agents/skills/; add -g for ~/.codex/skills/.

GitHub CLI

$ gh skill install iii-hq/workers judge --agent codex

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/iii-hq/workers.git skills-src && mkdir -p .agents/skills && cp -r skills-src/judge/skills .agents/skills/judge && rm -rf skills-src

Use ~/.agents/skills/ instead of .agents/skills for a personal install.

Codex skills documentation · loads skills from .agents/skills/

Paste into Codex

Install the "judge" agent skill from https://github.com/iii-hq/workers/tree/main/judge/skills into .agents/skills/judge/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "judge", then confirm the skill loads.

Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add iii-hq/workers --skill judge -a cursor

Project install goes to .agents/skills/; add -g for ~/.cursor/skills/.

GitHub CLI

$ gh skill install iii-hq/workers judge --agent cursor

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/iii-hq/workers.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/judge/skills .cursor/skills/judge && rm -rf skills-src

Use ~/.cursor/skills/ instead of .cursor/skills for a personal install.

Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/

Paste into Cursor

Install the "judge" agent skill from https://github.com/iii-hq/workers/tree/main/judge/skills into .cursor/skills/judge/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "judge", then confirm the skill loads.

Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

gemini skills install

$ gemini skills install https://github.com/iii-hq/workers.git --path judge/skills

--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.

skills CLI

$ npx skills add iii-hq/workers --skill judge -a gemini-cli

Project install goes to .agents/skills/; add -g for ~/.gemini/skills/.

GitHub CLI

$ gh skill install iii-hq/workers judge --agent gemini-cli

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/iii-hq/workers.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/judge/skills .gemini/skills/judge && rm -rf skills-src

Use ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.

Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/

Paste into Gemini CLI

Install the "judge" agent skill from https://github.com/iii-hq/workers/tree/main/judge/skills into .gemini/skills/judge/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "judge", then confirm the skill loads.

Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

GitHub CLI

$ gh skill install iii-hq/workers judge

Installs for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).

skills CLI

$ npx skills add iii-hq/workers --skill judge -a github-copilot

Project install goes to .agents/skills/; add -g for ~/.copilot/skills/.

Manual copy

$ git clone --depth 1 https://github.com/iii-hq/workers.git skills-src && mkdir -p .github/skills && cp -r skills-src/judge/skills .github/skills/judge && rm -rf skills-src

Use ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.

GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/

Paste into GitHub Copilot

Install the "judge" agent skill from https://github.com/iii-hq/workers/tree/main/judge/skills into .github/skills/judge/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "judge", then confirm the skill loads.

GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

skills CLI

$ npx skills add iii-hq/workers --skill judge -a opencode

OpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.

GitHub CLI

$ gh skill install iii-hq/workers judge --agent opencode

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/iii-hq/workers.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/judge/skills .opencode/skills/judge && rm -rf skills-src

Use ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.

OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/

Paste into OpenCode

Install the "judge" agent skill from https://github.com/iii-hq/workers/tree/main/judge/skills into .opencode/skills/judge/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "judge", then confirm the skill loads.

OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.

Facts

Skill name: judge
GitHub stars: 113
Token cost: ~1.5k tokens
SKILL.md length: 615 words
Files: 1
Skills in repo: 19
Repo updated: Oct 10, 2026
First seen: Oct 7, 2026
Licence: Apache-2.0

At a glance

Evaluate JSON state with Noul, Choice or Score questions for ticket triage, classification and graded judgments through the selected judge provider (TypeSafe JEV by default), or cancel an active…

Works in 3 steps: Read the registered contract with → For evaluation, supply 1–512 evaluations… → Invoke using the positional function…
Tasks that involve Customer support
SKILL.md covers Invocation, Cancel an active call, Resolve chat mentions and Result and operational…
Needs TYPESAFE_API_KEY

What it does

Judge is an agent skill from iii-hq/workers. Evaluate JSON state with Noul, Choice or Score questions for ticket triage, classification and graded judgments through the selected judge provider (TypeSafe JEV by default), or cancel an active request owned by the caller.

Its SKILL.md is about 1.5k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Sales & Support, covering Customer support. The licence is Apache-2.0.

When your agent uses it

Tasks that involve Customer support

Example prompts

“/judge”

Requirements

A credential in TYPESAFE_API_KEY

Workflow steps

3 steps, taken from the first numbered list in SKILL.md.

Read the registered contract with
For evaluation, supply 1–512 evaluations with unique nonblank IDs, JSON state
Invoke using the positional function name and --json, then inspect status

What it can do on your machine

Read from SKILL.md and the folder at commit ebfe027. It shows what the files ask for, not the result of running them.

Tool permissions
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Runs code
No scripts in the folder and no shell commands in SKILL.md (its code samples are bash).
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names these keys or tokens, usually read from environment variables:
- TYPESAFE_API_KEY
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Judge loads about 1.5k tokens when it runs. Until then it costs about 57 tokens; SKILL.md has 615 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it: ~57
When it runs · the whole SKILL.md, loaded when a task matches: ~1.5k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from iii-hq/workers at commit ebfe027, republished under its Apache-2.0 licence (© iii-hq). 615 words, ~1,470 tokens.

Download SKILL.mdSave it as .claude/skills/judge/SKILL.md (or your agent's skills folder).

name: judge
description: Evaluate JSON state with Noul, Choice or Score questions for ticket triage, classification and graded judgments through the selected judge provider (TypeSafe JEV by default), or cancel an active request owned by the caller.

judge

judge::evaluate evaluates supplied state with mixed Noul, Choice and Score questions. judge::cancel signals cancellation of an identified evaluation owned by the same caller. Each call is forwarded to judge-<provider>: the Default provider under Console Settings → Workers → judge (seeded from JUDGE_PROVIDER, default typesafe) or a top-level provider field selects it. Credentials and the default model live with the provider: for TypeSafe, configure api_key in the judge-typesafe entry under Console Settings → Workers or supply TYPESAFE_API_KEY to that worker process. An unregistered provider returns code: "provider_unavailable".

Invocation

Read the registered contract with iii trigger engine::functions::info --json '{"function_id":"judge::evaluate"}'. Use judge::cancel as function_id for that contract. For unfamiliar question/reply shapes, read the mixed request and typed response examples.
For evaluation, supply 1–512 evaluations with unique nonblank IDs, JSON state and a nonempty map of distinct question IDs. Choose primitives using the table below. Supply a positive timeout_ms within the operator's max_timeout_ms (default 300000). For queued work, also set expires_at_unix_ms, an absolute Unix deadline in milliseconds. Omit model to use the worker default. options.attempt_timeout_ms bounds each network attempt; retries are the provider's fixed policy, see Request options and retries.
Invoke using the positional function name and --json, then inspect status before reading results. Completion requires all requested typed answers or a reported failure; keep stats.usage_complete with the observed usage.

Primitive	Question criteria	Typed answer
`noul`	Optional `true` and/or `false` descriptions; omitted, null or empty criteria are accepted.	`noul` in `[0, 1]`.
`choice`	1–255 named options.	`choice`, `probabilities`, `confidence`.
`score`	2–10 ordered, non-null levels.	`score` in `[0, levels - 1]`, `probabilities`, `confidence`, `legend`.

For all primitives, instructions may be omitted or contain strings, objects, arrays or null. Noul and Choice descriptions accept those same forms. Score levels accept strings, objects or arrays; null Score levels are invalid (confirmed by a provider 422 and the SDK schema, despite the advanced guide). Structured values remain structured in Score legends. Noul can describe just one endpoint. Bare numbers and booleans are not top-level descriptions. A minimal mixed call:

bash

iii trigger judge::evaluate --json '{
  "timeout_ms": 3000,
  "evaluations": [{
    "id": "ticket-42",
    "state": {"ticket": "Production checkout is down for all customers."},
    "questions": {
      "urgent": {
        "type": "noul",
        "instructions": "Does this ticket describe an urgent production outage?"
      },
      "department": {"type": "choice", "criteria": {"billing": null, "technical": null}},
      "severity": {"type": "score", "criteria": ["Routine", {"impact": "Production blocked"}]}
    }
  }]
}'

On status: "ok", read results["ticket-42"].answers, selecting the field matching each answer's type. Choice probabilities use option IDs; Score probabilities and legends use zero-based index strings. Treat confidence as returned by the provider; it is not a caller-defined threshold or a formula to recompute.

Show full SKILL.md (241 more words)Show less

Cancel an active call

Start the evaluation with a top-level request_id (≤128 printable ASCII characters, unique among the caller's active calls), then invoke judge::cancel with {"request_id":"..."} from the same caller connection: ownership comes from the engine's _caller_worker_id, and separate CLI invocations may get different ids. cancelled: true means the signal was accepted, not that the provider rolled anything back; still await the original call, which answers code: "cancelled". Absent, finished or other callers' ids return cancelled: false. Details and the bus example: cancellation.

Resolve chat mentions

judge::mentions::resolve { "text": "…@kanban(id=\"KAN-12\")…" } (or { "mentions": [{ "name": "kanban", "id": "KAN-12" }] }) resolves worker-defined @<name>(id="…") mentions through their providers: each one's status (resolved, not_found, unknown_provider, error), one-line summary and the details call that returns the full item. Agents rarely need it — the judge's harness hook already resolves the mentions users write into a <mentions> block — but it verifies a new provider (ade/mentions) and serves surfaces outside the harness. See chat mentions.

Result and operational boundaries

Success carries the effective model, all results and stats; errors carry status: "error", code, optional http_status, a bounded and key-redacted provider_error, retry_after_ms and stats, with no partial answers. Bus invocation failures are separate. See results and failures.
usage counters are independently nullable; stats.usage_complete is false whenever any counter is unknown, so zero never means free.
Deadlines, retries, permits and byte limits are in limits and compatibility. Credentials stay in the provider's configuration or environment; requests never carry keys, provider URLs or extra headers.

Files

Just SKILL.md in judge/skills of iii-hq/workers.

Open the folder on GitHubat commit ebfe027

Compare with similar skills

Judge next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Judge compared with similar skills
Skill	Stars	Used in	Tokens	Auto-check	Licence	Repo updated
Judge this skilliii-hq/workers	113	—	~1.5k	Automated safety check: Pass	Apache-2.0	yesterday
Siftranknoperator/siftrank	225	—	~2.9k	Automated safety check: Pass	MIT	21 days ago
Supporthaacked/dotfiles	134	—	~2.9k	Automated safety check: Pass	None	yesterday
Customer Support Agentmastra-ai/mastra	29k	—	~2.2k	Automated safety check: Pass	Custom licence	today
Ticket TriageBevel-Software/Hexis	100	1 repos	~2.9k	Automated safety check: Pass	Apache-2.0	today
Add Agent Adaptersafedep/gryph	172	—	~679	Automated safety check: Pass	Apache-2.0	6 days ago

Similar skills

Siftrank
noperator/siftrank
Find needles in haystacks with SiftRank. An agent skill from noperator/siftrank.
225 GitHub stars~2.9k tokensUpdated 21 days ago
Sales & SupportAuto-check passed
Support
haacked/dotfiles
Support hero workflow — start a ticket investigation with auto-organized notes, find existing notes, or generate the weekly highlights log.
134 GitHub stars~2.9k tokensUpdated yesterday
Sales & SupportAuto-check passed
Customer Support Agent
mastra-ai/mastra
Authoring playbook for building agents that triage and reply to customer messages — support tickets, email inquiries, chat questions, refund requests, or product issues.
29k GitHub stars~2.2k tokensUpdated today
Sales & SupportAuto-check passed
Ticket Triage
Bevel-Software/Hexis
Triage and prioritize a support ticket or customer issue. An agent skill from Bevel-Software/Hexis.
100 GitHub starsUsed in 1 repo~2.9k tokens
Sales & SupportAuto-check passed
Add Agent Adapter
safedep/gryph
A skill your agent uses when adding support for a new AI coding agent to Gryph, or when changing how an existing agent adapter is wired.
172 GitHub stars~679 tokensUpdated 6 days ago
Sales & SupportAuto-check passed
Agent Support Matrix
jazzyalex/agent-sessions
Maintain Agent Sessions agent support matrix and JSON/JSONL parsing compatibility.
895 GitHub stars~587 tokensUpdated 2 days ago
Sales & SupportAuto-check passed

Questions about Judge

What does Judge do?

Evaluate JSON state with Noul, Choice or Score questions for ticket triage, classification and graded judgments through the selected judge provider (TypeSafe JEV by default), or cancel an active…. Judge is an agent skill from iii-hq/workers. Evaluate JSON state with Noul, Choice or Score questions for ticket triage, classification and graded judgments through the selected judge provider (TypeSafe JEV by default), or cancel an active request owned by the caller.

When should I use Judge?

Judge fits situations like: tasks that involve Customer support.

How do I install Judge in Claude Code?

Run `npx skills add iii-hq/workers --skill judge -a claude-code`. Or copy the skill folder (judge/skills in iii-hq/workers) into .claude/skills/judge in your project. Claude Code loads it when a task matches its description.

How do I install Judge in Codex?

Run `npx skills add iii-hq/workers --skill judge -a codex`. Or copy the skill folder (judge/skills in iii-hq/workers) into .agents/skills/judge in your project. Codex loads it when a task matches its description.

Can I use Judge in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add iii-hq/workers --skill judge -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/judge, .gemini/skills/judge, .github/skills/judge and .opencode/skills/judge in your project.

What does Judge need to run?

Going by SKILL.md and its folder, Judge needs credentials named TYPESAFE_API_KEY. Our summary lists: A credential in TYPESAFE_API_KEY.

Does Judge access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Judge safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Judge use?

Judge is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Judge use?

About 1.5k tokens (SKILL.md is roughly 5.9k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Judge?

Skills that share tags, products or a category with Judge: Siftrank (noperator/siftrank, 225 stars), Support (haacked/dotfiles, 134 stars), Customer Support Agent (mastra-ai/mastra, 29k stars) and Ticket Triage (Bevel-Software/Hexis, 100 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Judge?

iii-hq (a GitHub organization) maintains it in iii-hq/workers, which has 113 GitHub stars. The repository holds 19 skills in this directory. The repository was last updated on October 10, 2026.

Source: iii-hq/workers on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.

Judge

Install Judge

Facts

At a glance

What it does

When your agent uses it

Example prompts

Requirements

Workflow steps

What it can do on your machine

Tool permissions

Runs code

Network

Credentials

Context cost

Safety

SKILL.md

judge

Invocation

Cancel an active call

Resolve chat mentions

Result and operational boundaries

Files

Compare with similar skills

Similar skills

Siftrank

Support

Customer Support Agent

Ticket Triage

Add Agent Adapter

Agent Support Matrix

More from iii-hq/workers

Cron

Email

HTTP

Kanban

Pubsub

Release Sync

Questions about Judge

Install Judge

Facts

At a glance

What it does

When your agent uses it

Example prompts

Requirements

Workflow steps

What it can do on your machine

Tool permissions

Runs code

Network

Credentials

Context cost

Safety

SKILL.md

judge

Invocation

Cancel an active call

Resolve chat mentions

Result and operational boundaries

Files

Compare with similar skills

Similar skills

More from iii-hq/workers

Related

Questions about Judge