eval-evolution

A controlled workflow for improving a coding-agent skill or existing MCP tool through repeated tests against fixed evaluation evidence.

In plain words
What is it for?
Use it to prepare, run, review, and promote experiments targeting one skill or MCP component with defined limits for trials, tokens, cost, time, and retries.
Why use it?
It keeps optimization campaigns bounded and makes changes reviewable before they are explicitly promoted.

Skill for Claude CodeCodex

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 skills/miaodx/roboclaws/eval-evolution
Any agent
npx skills add MiaoDX/roboclaws --skill eval-evolution
Clone the repo
git clone --depth 1 https://github.com/MiaoDX/roboclaws

Made for: Claude Code, Codex.

Per session 28 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 733 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 $0.00028 $0.00733
Opus 5 $0.00014 $0.00367
Sonnet 5 $0.00006 $0.00147
Haiku 4.5 $0.00003 $0.00073

Measured 2d ago against content hash a976866b6b44, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

eval-evolution 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.

skills/eval-evolution/SKILL.md · 87 lines

How it starts

The opening of the file, as written. The whole thing — 87 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Eval Evolution

Use this Skill for maintainer-owned optimization campaigns driven by frozen eval evidence. roboclaws.evals remains the control plane. The optimizer and robot trials are distinct OpenAI Agents SDK agents; a provider profile selects their models and never changes the agent engine.

Prepare

Freeze an eval_evolution_campaign_v1 JSON manifest before the first optimizer call. Bind one target kind and target, the baseline commit and target digest, mutable paths, optimizer and robot provider/model identities, paired training suites, the orchestrator-only sealed confirmation reference, quality and minimum-improvement policy, runtime identity, and explicit turn/trial/token/ cost/time/retry ceilings. The budget block also declares positive optimizer_call_tokens, optimizer_call_cost_usd, robot_attempt_tokens, and robot_attempt_cost_usd reservations. Each reservation is a frozen maximum for one optimizer run or one robot attempt, not a post-hoc usage target.

Skill campaigns target exactly one skills/<name>/SKILL.md. Keep static-full as the baseline. no-skill is a non-promotable negative control. Do not mix Skill and MCP changes.

Run

Inspect the blocked preflight without provider execution:

just agent::eval evolve campaign=<campaign.json>

Run the frozen campaign only after provider/runtime readiness is established:

just agent::eval evolve campaign=<campaign.json> live_execution=run

The optimizer can only read the declared target, read sanitized feedback, and submit one hypothesis plus patch. It has no shell, filesystem, git, network, eval-launch, commit, or publication tool. Host-owned validation creates one content-addressed full baseline snapshot and runs deterministic gates before paired robot trials.

Review

Reject missing or mismatched identity, incomplete paired evidence, privacy or checker failures, trajectory/terminal violations, quality regressions, neutral rewrites, and no-skill. Efficiency ranks only candidates that meet the frozen quality and minimum-improvement rule. At most one training winner reaches one sealed confirmation. Its evidence never returns to the optimizer.

Read the full file on GitHub · 87 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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 · 87 lines · 28 tokens per session scan A a976866b6b44

Subscribe to this mod's changes

eval-evolution is a skill published in the GitHub repository MiaoDX/roboclaws (6 stars, last pushed 4d ago), licensed MIT. It adds 28 tokens to every session and 733 once invoked, about $0.0001 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.

Related

Other skills, from other repositories

zeroclaw

Help users operate and interact with their ZeroClaw agent instance — through both the CLI (zeroclaw commands) and the REST/WebSocket gateway API. Use this skill whenever the user wants to: send messages to ZeroClaw, manage memory or cron jobs, check system status, configure channels or providers, hit the gateway API…

zeroclaw-labs/zeroclaw · 167 tokens

add-dial-tool

Give chosen NanoClaw agents a real phone number as a container tool — the dial CLI baked into the agent image plus OneCLI credential injection for api.getdial.ai, scoped per agent, so the agents you pick can send SMS, place AI voice calls, and receive verification codes from inside the sandbox. Independent of the Dial…

nanocoai/nanoclaw · 117 tokens

slack-agent-flow

Let an existing Slack agent create new agents that arrive as their own Slack bots — provisioned app, operator DM, and a shared three-way room, hot-started without a host restart.

nanocoai/nanoclaw · 43 tokens

add-codex

Use Codex (OpenAI's codex app-server) as a full agent provider — planning, tool orchestration, MCP tools, server-side history, session resume — alongside or instead of Claude. ChatGPT subscription or OpenAI API key, vault-only via OneCLI. Per-group via ncl groups config update --provider codex. Distinct from using…

nanocoai/nanoclaw · 91 tokens

add-resend

Add Resend (email) channel integration via Chat SDK.

nanocoai/nanoclaw · 16 tokens

add-wechat

Add WeChat (personal) channel integration via Tencent's official iLink Bot API. Uses long-polling and QR scan — no webhook, no ToS risk, no paid token.

nanocoai/nanoclaw · 42 tokens