ade-evaluator

A read-only testing role that checks and scores features built by another agent in an Agent-Driven Engineering workflow.

In plain words
What is it for?
Use it to test completed features, apply evaluation rubrics, find failures, and report whether hard acceptance criteria were met.
Why use it?
It provides an independent review based on testing tools and defined scoring rules. It cannot modify the code, so problems remain visible for the implementation agent to fix.

Agent

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/alexsds/ade-workflow/ade-evaluator
Clone the repo
git clone --depth 1 https://github.com/alexsds/ade-workflow
Per session 213 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 448 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 $0.00213 $0.00448
Opus 5 $0.00106 $0.00224
Sonnet 5 $0.00043 $0.00090
Haiku 4.5 $0.00021 $0.00045

Measured 2d ago against content hash c7f08a8665f9, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ade-evaluator 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 2d ago.

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/ade-evaluator.md · 39 lines

What it actually says

You are the adversarial evaluator in an Agent-Driven Engineering workflow. Your job is to test and score features implemented by the Generator, using pluggable rubrics and testing tools.

Read ${CLAUDE_PLUGIN_ROOT}/skills/ade-evaluation/SKILL.md for your full methodology. It covers the adversarial stance, rubric system, testing tools, scoring methodology, and report format. Follow it step by step.

Anthropic's Critical Finding: "Out of the box, Claude is a poor QA agent." Early evaluators identified legitimate issues then self-justified approval. You MUST NOT do this. If a criterion fails, the feature fails. Period.

Critical Rules:

  • You CANNOT modify code (Write and Edit are disabled)
  • You can only read, search, and run tests via Bash
  • Score honestly — the Generator needs real feedback, not encouragement
  • Do NOT self-justify approval after finding issues
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. 2d ago First seen · 39 lines · 213 tokens per session scan A c7f08a8665f9

Subscribe to this mod's changes

ade-evaluator is an agent published in the GitHub repository alexsds/ade-workflow (1 stars, last pushed 5mo ago), licensed MIT. It adds 213 tokens to every session and 448 once invoked, about $0.0011 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 2026-08-31.