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 ericrisco/rsc-harness --skill finetuninggit clone --depth 1 https://github.com/ericrisco/rsc-harnessWrote 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/ericrisco/rsc-harness/finetuning)<a href="https://agentmods.dev/skills/ericrisco/rsc-harness/finetuning"><img src="https://agentmods.dev/badge/skills/ericrisco/rsc-harness/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/ericrisco/rsc-harness/finetuning"><img src="https://agentmods.dev/badge/skills/ericrisco/rsc-harness/finetuning.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00102 | $0.04070 |
| Opus 5 | $0.00051 | $0.02035 |
| Sonnet 5 | $0.00020 | $0.00814 |
| Haiku 4.5 | $0.00010 | $0.00407 |
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 7d 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 — 233 lines — stays where its author put it; the contents beside it link to each section on GitHub.
finetuning — teach an open model a form or behavior, not a fact
You own the discipline of adapting an open-weight model: deciding whether to fine-tune at all,
then running SFT and (optionally) preference optimization with trl + peft, backend-agnostic.
You are judged by whether the tuned model reliably produces the target form/behavior on a
held-out set — not by train loss, and not by vibes.
The one sentence that routes half of all "should I fine-tune?" questions correctly:
fine-tuning teaches form and behavior; RAG supplies facts. If the ask is "know our latest
prices / docs / tickets," that is retrieval (../rag/SKILL.md), not training. If the ask is "sound
like us, always emit this JSON, follow this reasoning pattern," that is here.
Decision gate — try this BEFORE reaching for a GPU
Fine-tuning is the last lever, not the first. Exhaust the cheaper, reversible options first; each row below is a real off-ramp.
| If the goal is… | Do this first | Fine-tune only when… |
|---|---|---|
| The model should know current/company facts | RAG (../rag/SKILL.md) — retrieve + ground |
never for facts; facts go stale, weights don't update |
| One-off format/tone, small volume | Prompt + few-shot (prompt-engineering) |
the prompt is huge, brittle, or you pay for it every call |
| Behavior depends on a long document | Longer context / put it in the prompt | context won't fit, or per-call token cost is the bottleneck |
| Consistent form/behavior at scale, latency/cost sensitive | — | prompting plateaus AND you have (or can build) good examples |
| A capability the base model just can't do | — | you have a reward signal or demonstration data for it |
Route out explicitly. Facts / freshness / citations → ../rag/SKILL.md. Squeezing a prompt before
spending money → prompt-engineering. Picking which base model (size/license/task) → open-weights.
Building the JSONL/preference corpus → training-data (LLM corpora, NOT tabular cleaning — that is
data-cleaning). A fast single-GPU run + GGUF export → ../unsloth/SKILL.md (same LoRA/QLoRA
concepts, one optimized implementation; this skill stays backend-agnostic). Downloading the base or
pushing the adapter/merged model → huggingface. Serving the result → ../vllm/SKILL.md.
What ships with it
4 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.
- 7d ago First seen · 233 lines · 102 tokens per session scan A 715a71d7b2aa
finetuning is a skill published in the GitHub repository ericrisco/rsc-harness (78 stars, last pushed yesterday), licensed MIT. It adds 102 tokens to every session and 4,070 once invoked, about $0.0005 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.
Other skills, from other repositories
fine-tuning
LLM fine-tuning authority — LoRA, QLoRA, and full fine-tuning workflows with PEFT, Axolotl, and Unsloth; supervised fine-tuning (SFT), DPO, and RLHF alignment; dataset curation and formatting; GPTQ/AWQ quantization; vLLM serving; and evaluation with lm-evaluation-harness.
fine-tuning-with-trl
Fine-tune LLMs using reinforcement learning with TRL - SFT for instruction tuning, DPO for preference alignment, PPO/GRPO for reward optimization, and reward model training. Use when need RLHF, align model with preferences, or train from human feedback. Works with HuggingFace Transformers.
axolotl
Expert guidance for fine-tuning LLMs with Axolotl - YAML configs, 100+ models, LoRA/QLoRA, DPO/KTO/ORPO/GRPO, multimodal support.
fine-tuning-with-trl
Fine-tune LLMs using reinforcement learning with TRL - SFT for instruction tuning, DPO for preference alignment, PPO/GRPO for reward optimization, and reward model training. Use when need RLHF, align model with preferences, or train from human feedback. Works with HuggingFace Transformers.
axolotl
Expert guidance for fine-tuning LLMs with Axolotl - YAML configs, 100+ models, LoRA/QLoRA, DPO/KTO/ORPO/GRPO, multimodal support.
axolotl
Expert guidance for fine-tuning LLMs with Axolotl - YAML configs, 100+ models, LoRA/QLoRA, DPO/KTO/ORPO/GRPO, multimodal support.