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 darrylmorley/ollama-plugin-cc --skill ollama-model-promptinggit clone --depth 1 https://github.com/darrylmorley/ollama-plugin-ccWrote 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/darrylmorley/ollama-plugin-cc/ollama-model-prompting)<a href="https://agentmods.dev/skills/darrylmorley/ollama-plugin-cc/ollama-model-prompting"><img src="https://agentmods.dev/badge/skills/darrylmorley/ollama-plugin-cc/ollama-model-prompting/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/darrylmorley/ollama-plugin-cc/ollama-model-prompting"><img src="https://agentmods.dev/badge/skills/darrylmorley/ollama-plugin-cc/ollama-model-prompting.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00023 | $0.01515 |
| Opus 5 | $0.00012 | $0.00758 |
| Sonnet 5 | $0.00005 | $0.00303 |
| Haiku 4.5 | $0.00002 | $0.00152 |
Grade A, and why
ollama-model-prompting 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 11d 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 — 100 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Ollama Model Prompting
Reference this skill when deciding which model to use and how to shape prompts for open-weight models.
Recommended Models Per Use Case
Empirically battle-tested. See docs/MODELS.md for the full results table, finding-count caveat, and reproducer.
| Use case | First choice | Notes |
|---|---|---|
| Cloud, anything | qwen3-coder-next:cloud |
Fastest tested anywhere (6–9 s per command). |
| Cloud, alt | glm-5.1:cloud |
Reliable structured output; clean rescue. |
| Local all-rounder | gpt-oss:20b |
Fastest local; reliable JSON; ~14 GB. |
| Local rescue | gemma4:26b |
Most resourceful when apply_patch rejects. |
| VRAM-constrained | qwen3.5:9b |
6.6 GB; works on every command. |
| Stop-review gate only | qwen3.5:9b or gpt-oss:20b |
Either handles the single-line ALLOW/BLOCK format. |
Select via --model <name> on any companion command. Falls back to OLLAMA_PLUGIN_DEFAULT_MODEL if set, otherwise the companion will error and prompt you to run /ollama:setup.
Models known to drift on the review JSON schema (use rescue-only or avoid): qwen3.6:27b-coding-nvfp4, batiai/qwen3.6-27b:q6, kimi-k2.6:cloud (review only — adversarial works).
Tool-Calling Support Matrix
Tool calling is required for the agentic rescue flow (default). Use --emit-patch to force patch-emit mode, which works without tool calling.
| Model family | Tool calling | Notes |
|---|---|---|
| Llama 3.1+ / Llama 4+ | Reliable | Native tool-call support since 3.1 |
| Llama 3.2 3B/1B | Unreliable | Too small; output format degrades |
| Qwen 2.5 / Qwen 3+ | Reliable | Solid tool-call format across sizes |
| Qwen 2.5/3 Coder | Reliable | Same base; code context does not hurt tool calls |
| DeepSeek-Coder-V2+ / DeepSeek-V2+ | Reliable | Strong reasoning; good tool adherence |
| DeepSeek-R1 (distills) | Unreliable | Thinking tokens interfere with JSON/tool output |
| Mistral 7B | Unreliable | v0.2 and earlier lack native tool-call format |
| Mistral Large / Nemo / Small | Reliable | Larger Mistral variants support tool calls |
| GPT-OSS (20B/120B) | Reliable | OpenAI open-weight; native tool-call format |
| Gemma 3+ | Reliable | Tool-call support added from Gemma 3 onward |
| Gemma 2 9B/27B | Partial | Tool-call-like output but not standard format |
| GLM 4+ | Reliable | Strong instruction-following and tool format |
| Kimi K2+ | Reliable | Cloud-hosted via Ollama; reliable tool calls |
| Command-R / Command-R+ | Reliable | Cohere; designed for tool use and RAG |
| Granite 3 | Reliable | IBM; native tool-call schema |
| Phi-3 / Phi-4 | Unreliable | Small; JSON adherence inconsistent |
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.
- 11d ago First seen · 100 lines · 23 tokens per session scan A 9d08c7fe3c20
ollama-model-prompting is a skill published in the GitHub repository darrylmorley/ollama-plugin-cc (4 stars, last pushed 4mo ago), licensed Apache-2.0. It adds 23 tokens to every session and 1,515 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.
Other skills, from other repositories
llm-app-patterns
Production-ready patterns for building LLM applications. Covers RAG pipelines, agent architectures, prompt IDEs, and LLMOps monitoring. Use when designing AI applications, implementing RAG, building agents, or setting up LLM observability.
prompt-optimization
Improve a prompt on the evaluations workbench through a measured loop. Score the baseline first, then duplicate the target column, form a hypothesis from failing rows, edit the copy's prompt draft, run, compare pass rate and cost, and repeat until the numbers hold. Use when the user asks to optimize or improve a…
enhance-prompt
Transforms vague UI ideas into polished, Stitch-optimized prompts. Enhances specificity, adds UI/UX keywords, injects design system context, and structures output for better generation results.
prompt-engineer
Writes, refactors, and evaluates prompts for LLMs — generating optimized prompt templates, structured output schemas, evaluation rubrics, and test suites. Use when designing prompts for new LLM applications, refactoring existing prompts for better accuracy or token efficiency, implementing chain-of-thought or few-shot…
seedance-vocab-en
This skill should be used when an English Seedance 2.0 prompt needs clearer production wording, less generic prose, or precise vocabulary for camera, lighting, motion, VFX, audio, and constraints. Route blocked prompts through seedance-filter for context and boundary review.
ideogram4
Prompting patterns for Ideogram 4 text-to-image — best-in-class in-image text rendering and exact color/layout control via structured JSON captions. Use when generating images that need legible on-image text (title cards, thumbnails, logos, signage, CTAs), precise brand colors, or controlled spatial layout. Triggers…