agentleFS
jakkarin-promseejakkarin-promsee/my-harness/.claude/skills/evaluate-model/SKILL.md

evaluate-model

Skill1 starsChanged 4 months agoMITView on GitHub

Evaluate a candidate model as the harness workhorse — pull it, configure .env, smoke-test, run the exit gate, compare token burn, record the verdict. Use when trying a new Ollama model or when the current workhorse misbehaves.

  • Reads credentials

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. gu-dsan6725/ai-assisted-coding★ 0same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository9 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Prefer instruct / non-thinking variants: thinking variants burned ~10× the output tokens here and died at length limits.

    line 13
  2. Very specific:

    If the candidate loses, revert .env.

    line 23

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Evaluate a model

This procedure has run three times (llama3.2:3b → qwen3:4b thinking → qwen3:4b-instruct won) and made the current call. Model evaluation is a core competency here — the whole "claude2 insurance" premise depends on being able to re-run this whenever the landscape shifts.

Hard constraints

  • 4 GB VRAM (RTX 3050 Ti Laptop) — the model must fit, or it swaps and crawls.
  • Prefer instruct / non-thinking variants: thinking variants burned ~10× the output tokens here and died at length limits.
  • Evaluate on local only. An unproven model driving the loop is exactly the "possibly-buggy loop" the budget rule bans from paid endpoints.

Steps

  1. ollama pull <model>; set OLLAMA_MODEL in .env; confirm OLLAMA_CONTEXT_LENGTH=8192 is in effect (Ollama's 4096 default silently truncates).
  2. Smoke: uv run harness ollama, one simple tool-using request ("read pyproject.toml and name the deps"). Watch for the two disqualifiers: does it call tools at all, and does it fabricate?
  3. Gate: uv run python acceptance/run.py ollama. Required: 7/7. Note the out-token count — the sitting champion (qwen3:4b-instruct) does the suite in ~1,460 out-tokens (2026-07-07); a candidate burning multiples of that loses even at 7/7.
  4. Quirk sweep: temperature behavior (never 0 on the qwen3 family), empty replies, cut-offs, tool-call JSON fidelity.
  5. Record the verdict in the root CLAUDE.md status line — winner or loser, with the one-line reason. The losers list is as valuable as the winner.
  6. If the candidate loses, revert .env.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.