Borrowing it
Nothing to install: this file belongs to nyosegawa/agentic-bench. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/nyosegawa/agentic-bench/main/.claude/skills/model-researcher/SKILL.mdgit clone --depth 1 https://github.com/nyosegawa/agentic-benchWrote 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/nyosegawa/agentic-bench/model-researcher)<a href="https://agentmods.dev/skills/nyosegawa/agentic-bench/model-researcher"><img src="https://agentmods.dev/badge/skills/nyosegawa/agentic-bench/model-researcher/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/nyosegawa/agentic-bench/model-researcher"><img src="https://agentmods.dev/badge/skills/nyosegawa/agentic-bench/model-researcher.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 | $0.00074 | $0.01484 |
| Opus 5 | $0.00037 | $0.00742 |
| Sonnet 5 | $0.00015 | $0.00297 |
| Haiku 4.5 | $0.00007 | $0.00148 |
Grade A, and why
model-researcher 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 10d 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 — 135 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Model Researcher
You are an experienced ML engineer investigating a model before running benchmarks.
Your Goal
Given a model name, produce a structured research summary:
- Model identity (type, architecture, parameter count, license)
- Hardware requirements (VRAM, recommended GPU)
- Recommended provider (cheapest viable option)
- Evaluation strategy (what to test and how)
Workflow
Step 1: Gather Model Information
Run the helper script to pull structured metadata:
python .claude/skills/model-researcher/scripts/hf_model_info.py MODEL_ID
To check if the model is available on HF Inference API (serverless):
python .claude/skills/model-researcher/scripts/hf_inference_check.py MODEL_ID
To search for similar or alternative models:
python .claude/skills/model-researcher/scripts/hf_model_search.py --task llm --sort downloads --limit 10
python .claude/skills/model-researcher/scripts/hf_model_search.py --search "qwen" --limit 5
If the script fails or the model is not on HuggingFace:
- Search the web for the model's official page, paper, or GitHub repo
- Manually gather: architecture, parameter count, input/output modalities, license
Step 2: Estimate VRAM and Cost
Run the estimator with --check-env to filter by available tokens and sort by cost:
python .claude/skills/model-researcher/scripts/gpu_estimator.py \
--params PARAM_COUNT --quant fp16 --model-type llm --check-env --json
--check-env reads .env and:
- Shows which providers have tokens set (✓/✗)
- Filters out providers without required tokens
- Sorts remaining by cheapest hourly rate
- Shows free tier / subscription info
Review the cost estimate and include it in the research summary. Adjust based on:
- Model-specific requirements (e.g., diffusion models need extra VRAM for image buffers)
- Framework overhead (transformers vs vllm vs diffusers)
Step 3: Determine Model Type and Evaluation Strategy
Classify the model and load the appropriate evaluation guide:
What ships with it
16 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.
- references/eval-3d-gen.md 2.4 KB
- references/eval-audio.md 2.2 KB
- references/eval-code-gen.md 2.9 KB
- references/eval-embedding.md 3.6 KB
- references/eval-image.md 1.6 KB
- references/eval-llm.md 1.7 KB
- references/eval-object-detection.md 3.3 KB
- references/eval-stt.md 2.6 KB
- references/eval-timeseries.md 2.0 KB
- references/eval-tts.md 1.9 KB
- references/eval-video-gen.md 2.5 KB
- references/eval-vlm.md 2.9 KB
- scripts/gpu_estimator.py 14 KB runs code
- scripts/hf_inference_check.py 4.0 KB runs code
- scripts/hf_model_info.py 4.9 KB runs code
- scripts/hf_model_search.py 4.2 KB runs code
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.
- 10d ago First seen · 135 lines · 74 tokens per session scan A 17d7ce18984d
model-researcher is a skill published in the GitHub repository nyosegawa/agentic-bench (5 stars, last pushed 6mo ago), licensed MIT. It adds 74 tokens to every session and 1,484 once invoked, about $0.0004 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
agent-platform-rag-engine-management
Manage and query Agent Platform RAG Engine Corpora and retrieve grounded contexts using the Google GenAI SDK. Use when listing RAG corpora or files, inspecting a corpus, retrieving contexts, or generating content grounded in a RAG corpus. Do not use for standard database queries (use SQL/Spanner skills), Google…
agent-platform-model-registry
Agent Platform Model Registry Management. Use when you need to upload, list, describe, update, or delete machine learning models (and their versions) in the Agent Platform Model Registry. Don't use for model training, model deployment to endpoints, or managing non-Agent Platform models.
foundry-config-setup
Resolve missing setup caused by a hardcoded Foundry project endpoint or model in a sample. Use when a sample fails because it uses a placeholder/hardcoded projectendpoint (for example "https://your-project.services.ai.azure.com") or a hardcoded model instead of reading them from the environment.
google-cloud-solution-agentic-analytics-spark-knowledge-catalog
Discovers requirements and generates guidance to design and deploy a governed, secure agentic-analytics solution for data that's distributed across Google Cloud, other cloud providers, or on-premises. Data that's outside Google Cloud (such as data from Databricks, Snowflake, Salesforce, SAP, or Oracle systems) is…
training-check
Interactively monitor training metrics from the current Codex session, periodically checking WandB or fallback logs for NaN, divergence, plateaus, and broken runs.
nemo-automodel-launcher-config
Configure NeMo AutoModel job launches for interactive runs, Slurm clusters, and SkyPilot cloud execution.