Evo is an agent plugin that turns a codebase into an automated experiment loop, discovering metrics, changing code, and testing whether those changes improve the results. It is used to optimize software through parallel subagents, tree-based exploration, shared experiment records, and optional regression or safety gates. The catalogue entries provide agent skills, hooks, commands, and other workflow components for operating Evo.
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.
npx skills add evo-hq/evo --skill finetuninggit clone --depth 1 https://github.com/evo-hq/evoWrote 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.
[](https://agentmods.dev/skills/evo-hq/evo/finetuning)<a href="https://agentmods.dev/skills/evo-hq/evo/finetuning"><img src="https://agentmods.dev/badge/skills/evo-hq/evo/finetuning/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/evo-hq/evo/finetuning"><img src="https://agentmods.dev/badge/skills/evo-hq/evo/finetuning.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Rogue Agent · line 137 Skill establishes unauthorized persistence across sessions via cron jobs, startup scripts, or state files. Session persistence allows an attacker to maintain access beyond the current interaction.Fix: Remove any persistence mechanisms (cron jobs, startup scripts, state files). Skills should not maintain state across sessions without explicit user consent.
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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00095 | $0.04106 |
| Opus 5.5 | $0.00038 | $0.01642 |
| Sonnet 5.5 | $0.00019 | $0.00821 |
| Haiku 4.5 | $0.00010 | $0.00411 |
Grade A, and why
finetuning 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 234 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Finetuning
Priors, not rules. Only firm guardrails: held-out eval you never train on, no leakage, trust evo's recorded numbers over the run's self-report. Override anything else against the gate.
Pick the technique by reward shape
Decide on the reward first, technique second. Choosing the comfortable technique over the matching one is the most common failure.
| Reward shape | Technique |
|---|---|
| Verifiable (exact match, unit tests, parser-decidable) | RL (GRPO / RLOO / PPO) — reward includes format, so the model learns to emit verifier-acceptable shape |
| Preference pairs (chosen vs rejected) | DPO / KTO / ORPO — cheaper than full RL, no rollouts |
| Demonstrations only (curated traces, chat data) | SFT — install format/tone/capability the base lacks |
| Have a scorer + want SFT stability | RFT — sample, filter by reward, SFT on survivors |
"SFT-then-RL" is not a law. For a competent base model on a verifiable benchmark, RL-from-base often beats SFT-then-RL end-to-end.
Research the literature before the first commit
The decision tree above is the structural prior. The empirical answer for this model on this benchmark usually has a recent paper, blog, or HF Space recipe behind it -- and what beats baseline on a 4B base model in 2026 is not what the agent's pre-training data captures. Before picking the technique for exp_0001 (the first experiment after baseline), invoke evo:ideator with a literature brief:
Task(
subagent_type="evo:ideator",
prompt="brief=literature\n"
"model_family=<e.g. Qwen3-4B-Base, Llama-3.1-8B-Base>\n"
"benchmark=<name + URL/paper if known>\n"
"objective=<one line: what beats baseline looks like>\n"
"constraints=<budget, data sources allowed, gated models forbidden, etc>"
)
The ideator returns ranked proposals with references (arXiv, HF Hub, GitHub, blogs). Read them before picking from the reward-shape table. A paper showing GRPO-from-base works on <model_family> for a similar verifiable benchmark beats applying the table cold.
What ships with it
8 files 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.
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.
- 1mo ago First seen · 234 lines · 95 tokens per session scan A ea19415079d8
finetuning is a skill published in the GitHub repository evo-hq/evo (1,486 stars, last pushed 22d ago), licensed Apache-2.0. It adds 95 tokens to every session and 4,106 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-08-30.
Other skills, from other repositories
codex-autoresearch
Triage improvement work and run or resume accepted measured loops in a local project. Architecture, documentation, UX, product study, open research, taste, and one-shot fixes stay direct unless the user explicitly requests repeated measurement with a complete experiment contract.
autoresearch-ml
Autonomous LLM training optimization with GPU support. Runs 5-minute training experiments, measures valbpb, keeps improvements or reverts — repeat forever. Use this skill when the user asks to "train a model autonomously", "optimize LLM training", "run ML experiments", "autoresearch with GPU", "optimize valbpb"…
recon-ai-enrichment
Wiring an LLM into a recon tool's decisions ("let AI pick {feature} for {tool}"): the never-raise contract, the per-target cache, the full+partial coverage, and the two UI toggles bound to one field. A raising AI helper or an empty fallback silently breaks or disables a live scan. Trigger: adding or editing a recon AI…
llm-provider-integration
Adding an LLM provider to RedAmon: the credential boundary (keys must never reach scan containers), prefix-routed model ids, and the provider registry. A misrouted model id or a leaked key are the two failures this guards. Trigger: adding or editing an LLM provider; editing parsemodelprovider in…
refactor-clean
Refactor cleanly instead of layering sediment. Use for rename requests, stale or misleading names, legacy leftovers, changed scope, or when a change reveals duplicated concepts, local adapters, obsolete owners, redundant inputs or guards, compatibility wrappers, parallel abstractions, an over-large module that has…
write-skills
Create or revise agent skills. Use when adding a new skill file, renaming a skill, simplifying an existing skill, improving trigger descriptions, or deciding what belongs in a skill versus references, scripts, assets, or ordinary docs.