Title: Orchestra-Research/AI-Research-SKILLs skills, scanned · ProSkills URL: https://proskills.dev/reports/repo/orchestra-research--ai-research-skills Description: Scan reports for 40 SKILL.md skills in Orchestra-Research/AI-Research-SKILLs: security findings, token cost and quality score, average 83/100. --- [Scan reports](https://proskills.dev/reports) · 13,703 stars # Orchestra-Research/AI-Research-SKILLs, scanned 40 skills from Orchestra-Research/AI-Research-SKILLs, average score 83/100, 3 with a critical finding. Scanned 2026-10-19. Best score first ## Every skill - [**distributed-llm-pretraining-torchtitan** Security 25/25 2,083 tokens Grade A 96/100](https://proskills.dev/reports/orchestra-research-distributed-llm-pretraining-torchtitan) - [**fine-tuning-with-trl** Security 25/25 2,715 tokens Grade A 96/100](https://proskills.dev/reports/orchestra-research-fine-tuning-with-trl) - [**implementing-llms-litgpt** Security 25/25 2,603 tokens Grade A 96/100](https://proskills.dev/reports/orchestra-research-implementing-llms-litgpt) - [**modal-serverless-gpu** Security 25/25 2,038 tokens Grade A 96/100](https://proskills.dev/reports/orchestra-research-modal-serverless-gpu) - [**nemo-curator** Security 25/25 2,165 tokens Grade A 96/100](https://proskills.dev/reports/orchestra-research-nemo-curator) - [**nnsight-remote-interpretability** Security 25/25 3,132 tokens Grade A 96/100](https://proskills.dev/reports/orchestra-research-nnsight-remote-interpretability) - [**peft-fine-tuning** Security 25/25 2,899 tokens Grade A 96/100](https://proskills.dev/reports/orchestra-research-peft-fine-tuning) - [**pytorch-lightning** Security 25/25 2,140 tokens Grade A 96/100](https://proskills.dev/reports/orchestra-research-pytorch-lightning) - [**pyvene-interventions** Security 25/25 3,403 tokens Grade A 96/100](https://proskills.dev/reports/orchestra-research-pyvene-interventions) - [**ray-data** Security 25/25 1,683 tokens Grade A 96/100](https://proskills.dev/reports/orchestra-research-ray-data) - [**simpo-training** Security 25/25 1,338 tokens Grade A 96/100](https://proskills.dev/reports/orchestra-research-simpo-training) - [**sparse-autoencoder-training** Security 25/25 3,042 tokens Grade A 96/100](https://proskills.dev/reports/orchestra-research-sparse-autoencoder-training) - [**transformer-lens-interpretability** Security 25/25 2,868 tokens Grade A 96/100](https://proskills.dev/reports/orchestra-research-transformer-lens-interpretability) - [**slime-rl-training** Security 25/25 2,648 tokens Grade A 92/100](https://proskills.dev/reports/orchestra-research-slime-rl-training) - [**training-llms-megatron** Security 25/25 2,262 tokens Grade A 92/100](https://proskills.dev/reports/orchestra-research-training-llms-megatron) - [**nanogpt** Security 25/25 1,551 tokens Grade A 90/100](https://proskills.dev/reports/orchestra-research-nanogpt) - [**rwkv-architecture** Security 25/25 1,631 tokens Grade A 90/100](https://proskills.dev/reports/orchestra-research-rwkv-architecture) - [**awq-quantization** Security 25/25 1,940 tokens Grade B 88/100](https://proskills.dev/reports/orchestra-research-awq-quantization) - [**huggingface-tokenizers** Security 25/25 3,251 tokens Grade B 88/100](https://proskills.dev/reports/orchestra-research-huggingface-tokenizers) - [**sentencepiece** Security 17/25 1,250 tokens Grade B 88/100](https://proskills.dev/reports/orchestra-research-sentencepiece) - [**pytorch-fsdp2** Security 25/25 2,606 tokens Grade B 87/100](https://proskills.dev/reports/orchestra-research-pytorch-fsdp2) - [**huggingface-accelerate** Security 25/25 1,946 tokens Grade B 86/100](https://proskills.dev/reports/orchestra-research-huggingface-accelerate) - [**openrlhf-training** Security 15/25 1,962 tokens Grade B 86/100](https://proskills.dev/reports/orchestra-research-openrlhf-training) - [**torchforge-rl-training** Security 15/25 2,368 tokens Grade B 82/100](https://proskills.dev/reports/orchestra-research-torchforge-rl-training) - [**unsloth** Security 25/25 479 tokens Grade B 82/100](https://proskills.dev/reports/orchestra-research-unsloth) - [**constitutional-ai** Security 25/25 1,907 tokens Grade B 80/100](https://proskills.dev/reports/orchestra-research-constitutional-ai) - [**miles-rl-training** Security 25/25 2,101 tokens Grade B 80/100](https://proskills.dev/reports/orchestra-research-miles-rl-training) - [**llamaguard** Security 25/25 2,163 tokens Grade C 78/100](https://proskills.dev/reports/orchestra-research-llamaguard) - [**prompt-guard** Security 13/25 2,235 tokens Grade C 78/100](https://proskills.dev/reports/orchestra-research-prompt-guard) - [**verl-rl-training** Security 15/25 2,322 tokens Grade C 78/100](https://proskills.dev/reports/orchestra-research-verl-rl-training) - [**ray-train** Security 13/25 2,524 tokens Grade C 76/100](https://proskills.dev/reports/orchestra-research-ray-train) - [**mamba-architecture** Security 25/25 1,696 tokens Grade C 72/100](https://proskills.dev/reports/orchestra-research-mamba-architecture) - [**llama-factory** Security 17/25 513 tokens Grade D 68/100](https://proskills.dev/reports/orchestra-research-llama-factory) - [**nemo-guardrails** Security 13/25 1,772 tokens Grade D 66/100](https://proskills.dev/reports/orchestra-research-nemo-guardrails) - [**lambda-labs-gpu-cloud** Security 0/25 2,920 tokens Grade F 63/100](https://proskills.dev/reports/orchestra-research-lambda-labs-gpu-cloud) - [**autoresearch** Security 3/25 5,982 tokens Grade D 60/100](https://proskills.dev/reports/orchestra-research-autoresearch) - [**deepspeed** Security 15/25 36,030 tokens Grade F 59/100](https://proskills.dev/reports/orchestra-research-deepspeed) - [**skypilot-multi-cloud-orchestration** Security 0/25 2,294 tokens Grade F 59/100](https://proskills.dev/reports/orchestra-research-skypilot-multi-cloud-orchestration) - [**grpo-rl-training** Security 0/25 4,196 tokens Grade F 57/100](https://proskills.dev/reports/orchestra-research-grpo-rl-training) - [**axolotl** Security 0/25 1,094 tokens Grade F 53/100](https://proskills.dev/reports/orchestra-research-axolotl) Source: [github.com/Orchestra-Research/AI-Research-SKILLs](https://github.com/Orchestra-Research/AI-Research-SKILLs) FAQ ## Questions **Is the skill's text copied here?** No. Each report keeps only the score, the findings and facts about the source. Get the skills themselves from Orchestra-Research/AI-Research-SKILLs on GitHub. **How current are these scores?** Scanned 2026-10-19. The list is rebuilt from GitHub on a schedule; paste a newer version into the scanner for a fresh score. Related ## Keep reading - [**Skill Scanner** Score any skill in your browser. Nothing is uploaded.](https://proskills.dev/scan) - [**SKILL.md format reference** Every field, its limit, the folder layout and the errors people hit.](https://proskills.dev/skill-md-format) - [**All guides** Every guide, grouped by what you are trying to do.](https://proskills.dev/guides) Free tool ## Scan your own skill Paste a SKILL.md or drop a folder. It is scored in your browser and nothing is uploaded. [Open the Skill Scanner](https://proskills.dev/scan) Want a skill that already passes? Start with a free one.