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 agentmods add skills/cxcscmu/skilllearnbench/nlp-project-setupnpx skills add cxcscmu/SkillLearnBench --skill nlp-project-setupgit clone --depth 1 https://github.com/cxcscmu/SkillLearnBenchWrote 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/cxcscmu/skilllearnbench/nlp-project-setup)<a href="https://agentmods.dev/skills/cxcscmu/skilllearnbench/nlp-project-setup"><img src="https://agentmods.dev/badge/skills/cxcscmu/skilllearnbench/nlp-project-setup.svg" alt="Measured on agentmods" 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 | $0.00017 | $0.00694 |
| Opus 5 | $0.00009 | $0.00347 |
| Sonnet 5 | $0.00003 | $0.00139 |
| Haiku 4.5 | $0.00002 | $0.00069 |
Grade A, and why
nlp-project-setup 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 4d 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 — 124 lines — stays where its author put it; the contents beside it link to each section on GitHub.
NLP Project Environment Setup
Environment Requirements for SimPO
Core Dependencies
- PyTorch: Deep learning framework (torch, torchvision, torchaudio)
- Transformers: Hugging Face library for LLMs
- NumPy: Numerical computing
- SciPy: Scientific computing utilities
- tqdm: Progress bars for training loops
Optional but Recommended
- wandb: Experiment tracking
- accelerate: Distributed training
- bitsandbytes: 8-bit optimization
- Flash-Attn: Efficient attention
Installation Steps
1. Check Python Version
python --version # Should be 3.8+
python -VV # Detailed version info
2. Create Virtual Environment (Optional)
python -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
3. Install Core Dependencies
# PyTorch (CUDA 12.1 example)
pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121
# Transformers
pip install transformers
# Other essentials
pip install numpy scipy tqdm
4. Verify Installation
python -c "import torch; print(torch.__version__)"
python -c "import transformers; print(transformers.__version__)"
Dependency Version Considerations
For SimPO Specifically
- transformers >= 4.30.0 (for AutoTokenizer, model loading)
- torch >= 1.13.0 (for modern PyTorch features)
- numpy (for .npz file saving)
Compatibility Notes
- Different CUDA versions may require different torch builds
- GPU memory requirements: typically 10-20GB for 7B models
- CPU-only mode works but is much slower
Requirements File
Create requirements.txt:
torch>=1.13.0
transformers>=4.30.0
numpy
scipy
tqdm
accelerate>=0.20.0
Then install:
pip install -r requirements.txt
Logging Installed Packages
# Save package list
python -m pip freeze > /root/python_info.txt
# Or capture with version info
python -VV >> /root/python_info.txt
python -m pip freeze >> /root/python_info.txt
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.
- 4d ago First seen · 124 lines · 17 tokens per session scan A 2f3bd0a35da4
nlp-project-setup is a skill published in the GitHub repository cxcscmu/SkillLearnBench (82 stars, last pushed 1mo ago), licensed MIT. It adds 17 tokens to every session and 694 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 2026-08-30.
Other skills, from other repositories
ceo-setup
One-time onboarding for the executive/manager commitment workflow — delegation-heavy, meeting prep, decision capture, morning and evening digests. Creates a commitments project and installs two dashboard widgets. After successful setup this skill is excluded from selection until the marker file is deleted.
content-creator-setup
One-time onboarding for the content creator workflow — content pipeline stages, trend expiration, cross-platform cascades, heavy idea parking. After successful setup this skill is excluded from selection until the marker file is deleted.
idea-parking
Park interesting ideas for later consideration, resurface them periodically, and promote to commitments when ready.
routing-subtour-elimination
Subtour-elimination methods for TSP, VRP, pickup/dropoff routing, and routing MIPs with binary arc variables. Use when route-continuity constraints may permit disconnected cycles and the model needs MTZ constraints, flow-based connectivity constraints, DFJ subset cuts, or lazy/iterative subtour cuts.
scip-opt
SCIP optimization with PySCIPOpt. Use when facing an optimization problem with an objective, hard constraints, soft penalties, integer decisions, routing, assignment, scheduling, allocation, packing, capacity, inventory, or service-level rules. Prefer modeling and solving the problem with PySCIPOpt when it is…
ebcdic-overpunch-decoding
Reference for the EBCDIC "overpunch" / zoned-decimal sign convention where the units position of a numeric field is replaced with a letter that encodes both a digit and a sign. Useful when reading mainframe-style fixed-length tapes whose amount fields appear as digits followed by a letter (e.g. "0000000000123D" or…