リサーチ・学習のスキル

Route DRL-Pytorch reinforcement-learning workflows for standalone PyTorch Q-learning, DQN-family, PPO, DDPG, TD3, SAC, Atari DQN, and Actor-Sharer-Learner scripts.

日本語の概要は準備中です。原文の説明を表示しています。

Use DreamerV3 for world-model reinforcement learning training, embodied environment dataflow, JAX model internals, result operations, and troubleshooting.

日本語の概要は準備中です。原文の説明を表示しています。

Use DreamerV2 2.2.0 for TensorFlow world-model reinforcement-learning training, custom Gym integration, typed experiment configuration, replay/checkpoint management, and JSONL/TensorBoard evaluation.

日本語の概要は準備中です。原文の説明を表示しています。

Route and use the repository's PyTorch reinforcement-learning examples across tabular control, DQN, on-policy actor-critic, and off-policy continuous control.

日本語の概要は準備中です。原文の説明を表示しています。

Use DeepCTR-Torch for PyTorch CTR/recommender feature columns, single-task models, DIN/DIEN sequence models, and multi-task learning workflows.

日本語の概要は準備中です。原文の説明を表示しています。

Guides DeepMotionEditing BVH animation, skeleton-aware motion retargeting, motion style transfer, and Blender visualization workflows for research use.

日本語の概要は準備中です。原文の説明を表示しています。

Operate Tongyi DeepResearch and its WebAgent family: ReAct inference setup, rollout validation, benchmark evaluation, and variant routing.

日本語の概要は準備中です。原文の説明を表示しています。

Use CleanRL's single-file reinforcement learning scripts, evaluation utilities, experiment operations, and repository maintenance workflows safely.

日本語の概要は準備中です。原文の説明を表示しています。

Routes Chainer workflows for training, export, distributed learning, ChainerX, and checkout maintenance.

日本語の概要は準備中です。原文の説明を表示しています。

Routes DisCo Researcher to Cream-family vision NAS, compression, distillation, and relative-position-encoding workflows across the AutoFormer, AutoFormerV2, Cream, CDARTS, EfficientViT, MiniViT, TinyCLIP, TinyViT, and iRPE project families.

日本語の概要は準備中です。原文の説明を表示しています。

Route auto-sklearn AutoML estimator, data validation, search, custom component, and metadata-maintenance workflows.

日本語の概要は準備中です。原文の説明を表示しています。

Use ARIS, the Auto Research in Sleep skill-based research harness, for setup, workflow routing, reviewer backend integration, state recovery, experiment operations, and repository maintenance.

日本語の概要は準備中です。原文の説明を表示しています。

Operate the Deep Researcher Agent repository for autonomous deep-learning experiments, execution backends, provider/tool dispatch, GPU resources, durable progress, and research-support integrations.

日本語の概要は準備中です。原文の説明を表示しています。

Routes AXLearn training, language-model, vision, audio/ASR, and GCP CLI workflows.

日本語の概要は準備中です。原文の説明を表示しています。

Guide agents through AgiBot X1 humanoid reinforcement-learning training, checkpoint playback, policy export, and MuJoCo sim2sim workflows with verified configuration contracts and explicit Isaac Gym limits.

日本語の概要は準備中です。原文の説明を表示しています。

Route AI-Optimizer reinforcement-learning collection tasks across model-based RL, easy-MARL, offline RL, safe command builders, and repository limitations.

日本語の概要は準備中です。原文の説明を表示しています。

Ingest a CVE writeup, blog post, advisory, or GitHub repo into the wiki - fetch, dedup via sources:, update the right technique/tool page(s), re-index. Generic knowledge only; never client data.

日本語の概要は準備中です。原文の説明を表示しています。

Drive responsible disclosure of a proven finding to a CVE. Package the report, find the vendor contact, report privately, coordinate a timeline, request the CVE (vendor CNA / GitHub / MITRE), and publish an advisory. Closes the research loop. Triggers - "disclose", "request a cve", "report this to the vendor".

日本語の概要は準備中です。原文の説明を表示しています。

Vulnerability-research loop toward a novel CVE. Target triage -> attack-surface map -> ranked hypotheses -> investigate (RE / fuzz / audit) -> a finding deepens the loop, a dead-end pivots to a new approach. Uses the full wiki + hunt skillset. Scaffolds and persists state under raw/research/<project>/. Triggers - "research", "find a cve", "analyze this binary/library", "audit this code for vulns".

日本語の概要は準備中です。原文の説明を表示しています。

Post-engagement knowledge harvest AND harness retrospective - after a box/bugbounty/pentest/CTF is completed, first diff how the engagement was EXECUTED against the skills/hooks that governed it (what discipline was skipped) and improve the harness, then sweep the whole engagement for GENERIC reusable knowledge NOT already in wiki/ and land it via the leak-gated stage->promote pipeline. Use at close-out or when asked to "extract learnings", "what did we learn", "harvest lessons into wiki", "distill this engagement", "post-mortem into the wiki", "what did we do wrong", "improve the harness from this box".

日本語の概要は準備中です。原文の説明を表示しています。

Use when users ask to 找文献、做文献检索、查论文、查临床试验、核验引用、去重文献、设计 PubMed/MeSH 检索式、追踪上下游引文、解析 DOI/PMID/PMCID/arXiv/OpenAlex/Semantic Scholar/NCT ID， 或导出 RIS、BibTeX、NBIB、ENW；also use for multi-source academic search, citation verification, citation graphs, trial registration search, research workflows, or DeepSeek Harness academic-search MCP setup.

日本語の概要は準備中です。原文の説明を表示しています。

Research and advise on the best way to approach a software project, including architecture, tech stack, implementation strategy, pricing/operating-cost tradeoffs, benchmark research, and comparisons with similar real-world projects. Use before building, mid-build, or after completion when the user asks for project strategy, optimal approach, research comparables, similar projects, stack selection, vendor/service choice, repo analysis, architecture critique, implementation feedback, or a prioritized improvement plan. Avoid for narrow single-bug debugging or isolated file edits unless the user asks for broader project direction.

日本語の概要は準備中です。原文の説明を表示しています。

Reconstruct and report long-running or multi-turn research, architecture questions, reviews, decisions, completion results, and status as a clear, self-contained brief. Use when a reader must re-enter earlier context, understand the real question and its relationships, distinguish verified facts from judgment and open evidence, or make a decision without reconstructing the work log. Quote the relevant original user wording before substantive analysis, keep one coherent main line, use spatial diagrams when relationships matter, and finish with an explicit decision boundary and practical next step.

日本語の概要は準備中です。原文の説明を表示しています。

Deeply research technical architecture, source code, mechanisms, SDKs, frameworks, project comparisons, and system-design options across repositories, history, official docs, issues, discussions, and high-quality community signals. Use when the user asks for deep research, source-level understanding, architecture comparison, mechanism satisfaction, or evidence for designing a system. Always map the question and obtain confirmation before beginning substantive research; skip for simple factual lookups or straightforward implementation tasks.

日本語の概要は準備中です。原文の説明を表示しています。