eval-judge

eval-judge is an agent for coding agents from Kanevry/session-orchestrator. It costs 249 tokens per session (1,959 once invoked), scanned A, original, MIT.

A read-only judge that evaluates a session record against two predefined criteria: following instructions and producing a useful report. It uses the record's evidence and metrics and returns advisory judgments without creating an overall score.

In plain words
What is it for?
Use it during an evaluation workflow to assess instruction adherence and report quality from a session-evaluation record.
Why use it?
It adds a consistent review of whether an evaluation session followed its instructions and documented its results honestly and specifically.

Agent

Part of the session-orchestrator plugin — 49 skills, 28 commands, 19 agents, 11 hooks, 1 MCP server shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/kanevry/session-orchestrator/eval-judge
Clone the repo
git clone --depth 1 https://github.com/Kanevry/session-orchestrator

Or install session-orchestrator, the plugin that ships this one along with the rest of its 49 skills, 28 commands, 19 agents, 11 hooks, 1 MCP server.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for eval-judge

README.md
[![agentmods](https://agentmods.dev/badge/agents/kanevry/session-orchestrator/eval-judge.svg)](https://agentmods.dev/agents/kanevry/session-orchestrator/eval-judge)
Your own site
<a href="https://agentmods.dev/agents/kanevry/session-orchestrator/eval-judge"><img src="https://agentmods.dev/badge/agents/kanevry/session-orchestrator/eval-judge.svg" alt="Measured on agentmods" height="20"></a>
Per session 249 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,959 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00249 $0.01959
Opus 5 $0.00125 $0.00979
Sonnet 5 $0.00050 $0.00392
Haiku 4.5 $0.00025 $0.00196

Measured yesterday against content hash 693ae845aaf5, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

eval-judge scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agents/eval-judge.md · 147 lines

How it starts

The opening of the file, as written. The whole thing — 147 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Eval-Judge Agent

You judge, from a session-eval record slice, whether the session showed instruction-adherence and whether the record's report-quality is honest and specific — the two pre-registered judge dimensions defined in skills/eval/rubric-v1.md § "Judge Dimensions" for the aiat-llm-eval/1.0 standard. You are dispatched by scripts/lib/eval/judge.mjs::runEvalJudge with a complete prompt — your job is to read the record slice, answer the two judge questions, and emit ONE fenced json block of exactly two judgment objects.

Your output is advisory only and always uncalibrated. It is merged into the session-eval record by the coordinator via mergeJudgeDimensions() and appended to .orchestrator/metrics/eval.jsonl via appendEvalRecord(). Per the standard's "no global score, by construction" rule, your judgments are never blended into the deterministic five-dimension tally and never produce or feed a global/overall score — they are visibly separated, advisory verdicts a reader can discard and still have a complete deterministic evaluation.

Color rationale (AGENTS.md exception (b) — mutually-exclusive phase): this agent carries color: cyan, shared with dialectic-deriver (/evolve phase), docs-writer (impl/finalization phase), and skill-applied-judge (session-end Phase 3.6.6). This judge runs solo, dispatched coordinator-side during the /eval skill's Phase 3, and never co-runs in a dispatch wave, so the shared cyan can never collide on screen.

Core responsibilities

  1. Judge instruction-adherence: from the record slice, decide whether the coordinator appears to have followed the operator's stated instructions and the repo's always-on rules (verification-before-completion, ask-via-tool, parallel-session safety, scope discipline) — pass, fail, not-applicable, or cannot-determine when the slice gives no clear signal.
  2. Judge report-quality: decide whether the record's evidence reads as honest, specific, and evidence-anchored (no "should pass" without a run, no superlatives, drift/carryover named plainly) versus vague, self-congratulatory, or padded — same four-state verdict.
  3. Never guess: prefer cannot-determine over a confident guess when the record slice is silent or ambiguous on a question. A missing signal is not evidence either way.
  4. Stay in scope: emit exactly one judgment per dimension in the fixed set (instruction-adherence, report-quality) — never invent a third dimension, never omit one of the two.

Read the full file on GitHub · 147 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 147 lines · 249 tokens per session scan A 693ae845aaf5

Subscribe to this mod's changes

eval-judge is an agent published in the GitHub repository Kanevry/session-orchestrator (49 stars, last pushed yesterday), licensed MIT. It adds 249 tokens to every session and 1,959 once invoked, about $0.0012 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.

Related

Other agents, from other repositories

providers

A coding agent runs against a provider account: a login, on a plan, that may or may not be metered. That account has a two-tier balance: included subscription windows that refill on their own clocks, plus paid extra or API usage the provider or the local spend store can name.

rimio-ai/rimz · 0 tokens

model

A coding agent runs in a pane, reports through its own hooks, and appears in the sidebar as one card. This doc owns the model in between: how a native event becomes one durable state per agent, how that state moves, and how the row you see is projected from it.

rimio-ai/rimz · 0 tokens

adapter

Thirteen coding agents report to RimZ, and no code outside crates/rimz/src/agents/ knows which one it is looking at. This doc owns the seam that makes that true: what an adapter is, the contracts it implements, and the path a native hook event walks from the agent's process to the durable store.

rimio-ai/rimz · 0 tokens

adapter_opencode

This doc is the single home for everything OpenCode-specific. OpenCode's integration surface is an in-process TypeScript plugin, so the adapter ships plugin.ts, embedded at compile time and installed as /.config/opencode/plugin/rimz.ts. The plugin subscribes to OpenCode hooks and bus events, flattens them into a…

rimio-ai/rimz · 0 tokens

adapter_pi

This doc is the single home for everything Pi-specific. Pi's integration surface is in-process TypeScript extensions, so the adapter ships one — extension.ts, embedded at compile time — that forwards each event below to rimz hooks feed --source pi as a fire-and-forget child. The child direction inverts: Claude and…

rimio-ai/rimz · 0 tokens

adapter_copilot

GitHub Copilot CLI reports through native camelCase command hooks installed as one RimZ-owned user file at $COPILOTHOME/hooks/rimz.json, falling back to /.copilot/hooks/rimz.json. Each command passes its event name to rimz hooks feed --source copilot --event because native payloads carry no event-name field. Hook…

rimio-ai/rimz · 0 tokens