一个技能，一个新的可能。找到适合你的 AI 工作流。
llm-distillation
【数据与研究】Distill knowledge from large teacher LLMs into small student models via sequence-level (response-based) knowledge distillation. Covers teacher selection and licensing checks, generating distillation data with Bedrock serverless teachers (DeepSeek-R1, Claude via converse), reasoning distillation (keeping <think>/chain-of-thought for math/code/puzzle d…
academic-course-setup-automator
【设计与创意】When the user needs to set up multiple academic courses in a learning management system (Canvas/LMS) from structured data sources. This skill automates the entire workflow extracting course schedules from emails/attachments, matching instructors from CSV files, creating courses, enrolling teachers, publishing announcements with class details, uploadi…
academic-course-setup-automator
【设计与创意】When the user needs to set up multiple academic courses in a learning management system (Canvas/LMS) from structured data sources. This skill automates the entire workflow extracting course schedules from emails/attachments, matching instructors from CSV files, creating courses, enrolling teachers, publishing announcements with class details, uploadi…
academic-course-setup-automator
【设计与创意】When the user needs to set up multiple academic courses in a learning management system (Canvas/LMS) from structured data sources. This skill automates the entire workflow extracting course schedules from emails/attachments, matching instructors from CSV files, creating courses, enrolling teachers, publishing announcements with class details, uploadi…
curriculum-designer
【文档与办公】Helps teachers and instructional designers create a structured curriculum or teaching plan from a list of topics or a specification document. Use when the user wants to design course materials, lesson plans, or a syllabus.
d-core-task-decomposition
【效率与协作】Mitigate lazy reasoning in Large Reasoning Models via self-distillation teaching task decomposition, followed by Diversity-Aware GRPO with entropy-based advantage functions, enabling effective decomposition without external teachers while balancing structured reasoning with diversity.
google-classroom
【开发与工程】Google Classroom API integration with managed OAuth. Manage courses, assignments, students, teachers, and announcements. Use this skill when users want to create courses, manage coursework, track student submissions, or post announcements. For other third party apps, use the api-gateway skill (https://clawhub.ai/byungkyu/api-gateway). Requires networ…
gtr-turbo-vlm
【效率与协作】Eliminate expensive external teacher dependencies in VLM RL training via merged-checkpoint teachers. Uses TIES merging of historical RL checkpoints to create free, stable teacher models for step-level guidance—matching external teacher performance while reducing training time 50% and computational costs 60%.
learning-beyond-teacher
【效率与协作】Train student models to exceed teacher performance by extrapolating rewards during distillation using a scaling factor λ1. Enables unified students to surpass all individual domain teachers through controlled reward amplification.
let-late-early-training
【效率与协作】Accelerate LLM pretraining by leveraging small pretrained models as teachers, enabling large models to explicitly learn late-layer knowledge in early layers via alignment loss, achieving 1.6x convergence speedup and 5% downstream improvement even with 10x smaller teachers.
pmla-topic-selection
【开发与工程】Use when deciding whether a literary or language-studies project fits PMLA (Publications of the Modern Language Association) and which venue to target. PMLA is the discipline's generalist flagship, so the test is interest to scholars and teachers of language and literature broadly, not narrow specialist novelty. Helps frame the question; it does not …
remit-rl-guided-mid-training
【数据与研究】Improve LLM reasoning by reweighting pre-training data during mid-training based on discrepancies between RL-tuned and base models, boosting reasoning performance without external teachers or extra data.
self-eval-t2i
【效率与协作】Train text-to-image models from scratch for any inference step count via self-evaluation mechanism. Model evaluates its own generated samples using current score estimates as dynamic self-teacher, enabling global distribution matching without external teachers—achieving few-step quality equivalent to many-step models at all budgets.
wps-courseware
【设计与创意】教学课件制作。老师和培训师的做课件助手，按照"导入→新授→练习→总结"的 教学设计结构自动生成课件PPT，帮你拆解知识点、设计互动环节、生成练习题。 用于帮助教师快速制作课件。当用户提到课件、教学PPT、培训材料时触发。 Courseware maker for teachers with structured lesson design.