hermes-agent
github.com/NousResearch/hermes-agent
| Skill | Added | Review |
|---|---|---|
axolotl optional-skills/mlops/training/axolotl/SKILL.md Axolotl: YAML LLM fine-tuning (LoRA, DPO, GRPO). | 47 47 Impact — No eval scenarios have been run Securityby  Low Low-risk findings worth noting Version: 6680afb | |
torchtitan optional-skills/mlops/torchtitan/SKILL.md Pretrain LLMs at scale with PyTorch 4D parallelism. | 56 56 Impact — No eval scenarios have been run Securityby High Do not use without reviewing Version: 6680afb | |
tensorrt-llm optional-skills/mlops/tensorrt-llm/SKILL.md High-throughput LLM inference on NVIDIA GPUs. | 48 48 Impact — No eval scenarios have been run Securityby Low Low-risk findings worth noting Version: 6680afb | |
stable-diffusion optional-skills/mlops/stable-diffusion/SKILL.md Text-to-image generation, inpainting, and img2img. | 56 56 Impact — No eval scenarios have been run Securityby Critical Do not install without reviewing Version: 6680afb | |
slime optional-skills/mlops/slime/SKILL.md RL post-training for LLMs with Megatron and SGLang. | 48 48 Impact — No eval scenarios have been run Securityby Low Low-risk findings worth noting Version: 6680afb | |
simpo optional-skills/mlops/simpo/SKILL.md Reference-free preference alignment, simpler than DPO. | 55 55 Impact — No eval scenarios have been run Securityby Low Low-risk findings worth noting Version: 6680afb | |
saelens optional-skills/mlops/saelens/SKILL.md Train sparse autoencoders to interpret model features. | 55 55 Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: 6680afb | |
dspy optional-skills/mlops/research/dspy/SKILL.md DSPy: declarative LM programs, auto-optimize prompts, RAG. | 49 49 Impact — No eval scenarios have been run Securityby Critical Do not install without reviewing Version: 6680afb | |
qdrant optional-skills/mlops/qdrant/SKILL.md Vector search engine for production RAG systems. | 53 53 Impact — No eval scenarios have been run Securityby High Do not use without reviewing Version: 6680afb | |
pytorch-lightning optional-skills/mlops/pytorch-lightning/SKILL.md Clean training loops with built-in distributed support. | 52 52 Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: 6680afb | |
pytorch-fsdp optional-skills/mlops/pytorch-fsdp/SKILL.md Fully sharded data-parallel training for large models. | 44 44 Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: 6680afb | |
pinecone optional-skills/mlops/pinecone/SKILL.md Managed vector DB for production RAG and search. | 50 50 Impact — No eval scenarios have been run Securityby High Do not use without reviewing Version: 6680afb | |
peft optional-skills/mlops/peft/SKILL.md Fine-tune large LLMs with LoRA on limited GPU memory. | 56 56 Impact — No eval scenarios have been run Securityby Low Low-risk findings worth noting Version: 6680afb | |
obliteratus optional-skills/mlops/obliteratus/SKILL.md OBLITERATUS: abliterate LLM refusals (diff-in-means). | 51 51 Impact — No eval scenarios have been run Securityby Critical Do not install without reviewing Version: 6680afb | |
nemo-curator optional-skills/mlops/nemo-curator/SKILL.md Curate LLM training data: dedupe, filter, PII redaction. | 52 52 Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: 6680afb | |
modal optional-skills/mlops/modal/SKILL.md Serverless GPU cloud for ML jobs and model APIs. | 56 56 Impact — No eval scenarios have been run Securityby High Do not use without reviewing Version: 6680afb | |
llava optional-skills/mlops/llava/SKILL.md Vision-language chat: VQA, captioning, image dialogue. | 48 48 Impact — No eval scenarios have been run Securityby Low Low-risk findings worth noting Version: 6680afb | |
lambda-labs optional-skills/mlops/lambda-labs/SKILL.md On-demand GPU cloud instances for ML training. | 46 46 Impact — No eval scenarios have been run Securityby Low Low-risk findings worth noting Version: 6680afb | |
instructor optional-skills/mlops/instructor/SKILL.md Structured LLM outputs validated with Pydantic. | 48 48 Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: 6680afb | |
outlines optional-skills/mlops/inference/outlines/SKILL.md Outlines: structured JSON/regex/Pydantic LLM generation. | 57 57 Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: 6680afb | |
huggingface-tokenizers optional-skills/mlops/huggingface-tokenizers/SKILL.md Fast BPE/WordPiece tokenization and custom vocab training. | 59 59 Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: 6680afb | |
guidance optional-skills/mlops/guidance/SKILL.md Constrain LLM output with grammars; guarantee valid JSON. | 48 48 Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: 6680afb | |
flash-attention optional-skills/mlops/flash-attention/SKILL.md Speed up long-sequence transformer training and inference. | 57 57 Impact — No eval scenarios have been run Securityby Low Low-risk findings worth noting Version: 6680afb | |
faiss optional-skills/mlops/faiss/SKILL.md Fast vector similarity search at billion scale. | 48 48 Impact — No eval scenarios have been run Securityby Passed No findings from the security scan Version: 6680afb | |
clip optional-skills/mlops/clip/SKILL.md Zero-shot image classification and image-text search. | 54 54 Impact — No eval scenarios have been run Securityby Low Low-risk findings worth noting Version: 6680afb |