A policy for deciding whether a task should use a local language model or the current agent. Local models run on the user's computer, while the current agent handles work needing more judgment.

Use it to route batch summaries, data cleanup, boilerplate, draft translations, and other mechanical tasks to a local model, while keeping reviews, design choices, and hard debugging with the agent.

It can reduce external model use for repetitive or sensitive work while keeping architectural decisions, difficult debugging, and final writing with the main agent.

local-model — Skill for Claude Code and Codex by fthsrbst: A policy for deciding whether a task should use…

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/fthsrbst/mnema/local-model

Any agent

npx skills add fthsrbst/mnema --skill local-model

Clone the repo

git clone --depth 1 https://github.com/fthsrbst/mnema

Made for: Claude Code, Codex.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/fthsrbst/mnema/local-model.svg)](https://agentmods.dev/skills/fthsrbst/mnema/local-model)

Your own site

<a href="https://agentmods.dev/skills/fthsrbst/mnema/local-model"><img src="https://agentmods.dev/badge/skills/fthsrbst/mnema/local-model.svg" alt="Measured on agentmods" height="20"></a>

Per session 60 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 685 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.

Model	Per session	Once invoked
Fable 5	$0.00060	$0.00685
Opus 5	$0.00030	$0.00342
Sonnet 5	$0.00012	$0.00137
Haiku 4.5	$0.00006	$0.00068

Measured 5d ago against content hash f2f9eef276af, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

local-model 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 5d 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/local-model/SKILL.md · 37 lines

What it actually says

Model Yönlendirme: Zorluğa Göre Karar

Temel kural: işin zorluğuna göre karar ver. Üç seviye:

Basit/toplu/mekanik → local_llm (LM Studio veya Ollama, ücretsiz, veri lokalde)
Orta ve zor → sen (mevcut agent) yaparsın — sen zaten güçlü bir bulut modelisin; ekstra API key veya entegrasyon gerekmez. Fatih'in kullandığı agent (Claude Code, opencode go aboneliği vb.) hangisiyse zor işin modeli odur.
Görsel/video/medya → media_generate (ComfyUI, [[media-gen]] skill'i)

Hub'a ayrıca bulut LLM API'si bağlamıyoruz — bilinçli karar: agent'ın kendisi o katman.

local_llm'e uygun işler

Toplu/tekrarlı işler: 50 dosyaya özet, commit mesajı üretimi, veri temizleme/dönüştürme, etiketleme
Taslaklar: ilk sürüm docstring, test iskeleti, çeviri taslağı (sonra sen rafine et)
Hassas veri: kullanıcının dışarı çıkmasını istemediği içerik üzerinde işlem
Kod tamamlama tarzı mekanik üretim: boilerplate, regex, SQL taslağı

Sende kalması gereken işler (local_llm'e YOLLAMA)

Mimari kararlar, kod review, zor debug — yargı gerektiren her şey
Kullanıcıyla diyalog ve nihai çıktı kalitesinin önemli olduğu metinler
Hub hafızasına yazılacak kayıtların son hali (kalite senin sorumluluğun)

Kullanım

machine_status → hangi makinede LM Studio/Ollama açık, hangi modeller var.
Model seçimi (isimden çıkarım):
- *coder* → kod işleri
- küçük model (7-20B) → hızlı basit işler; büyük (30B+) → kalite gereken taslaklar
- model belirtmezsen yüklü ilk model kullanılır — toplu işte önce 1 örnekle dene, çıktı iyiyse devam et.
local_llm(prompt, model?, backend?, system?, max_tokens?) — backend boşsa LM Studio öncelikli, backend: "ollama" ile Ollama'ya yönlendir. Uzun işlerde çıktıyı doğrula: yerel modeller talimat kaçırabilir; kritik alanları örneklemle kontrol et.

Kural

Sonucu kendi çıktın gibi sunma — yerel model çıktısını kullandıysan doğrulamandan geçir. Hata bulursan kendin düzelt, kullanıcıya ham yerel model çıktısı gösterme.

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.

5d ago First seen · 37 lines · 60 tokens per session scan A f2f9eef276af

Subscribe to this mod's changes

local-model is a skill published in the GitHub repository fthsrbst/mnema (2 stars, last pushed 5d ago), licensed MIT. It adds 60 tokens to every session and 685 once invoked, about $0.0003 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.

Data and AI ai-agents claude cli codex local-first mcp

esmfold2

Biohub ESMFold2 / ESMFold2-Fast all-atom co-folding (Candido et al. 2026, github.com/Biohub/esm). Single-sequence and MSA modes; protein, DNA, RNA, ligand (CCD/SMILES), modified residues. FoldBench Ab-Ag 50-55%, PPI 70-77% DockQ-pass. Also covers the ESMC-{300M,600M,6B} protein language models from the same release…

aipoch/open-science · 223 tokens

evo2

Score, embed, and generate DNA sequences with Evo 2, a long-context genomic foundation model. Use this skill when: (1) Computing per-nucleotide or per-sequence likelihoods for variant effect scoring, (2) Embedding genomic windows for downstream classification, (3) Generating DNA conditioned on a prefix, (4) Scoring…

aipoch/open-science · 83 tokens

scgpt

Embed and annotate single-cell expression data with scGPT, a foundation model for single-cell biology. Use this skill when: (1) Producing cell embeddings from an AnnData for clustering/integration, (2) Zero-shot or fine-tuned cell-type annotation, (3) Gene-level representation for perturbation/GRN tasks. For…

aipoch/open-science · 89 tokens

local-model