Loading...
Loading...
Found 159 Skills
Design hypothesis-driven ML/AI experiments before running them. Use this skill whenever the user wants to plan experiments, ablations, baselines, metrics, controls, seeds, logging, stop conditions, reviewer-proof evidence, or an experiment matrix for a paper claim before using run-experiment or writing results.
Initialize a full ML research project control root with independent paper, code, and optional slide repositories, shared project memory, root-level agent guidance, code-owned worktree policy, and component handoffs. Use when starting a new research project, setting up a project root for agents, connecting paper/code/slides repos, or replacing a simple paper+code workspace with a lifecycle-aware research project structure.
Use when testing, reviewing, pressure-testing, refining, packaging, or validating agent skills for academic research workflows before installing or relying on them.
Use when inspecting, cleaning, understanding, reproducing, or auditing academic research code repositories, especially when README commands, datasets, checkpoints, experiments, or paper claims need verification.
Use when academic research involves human subjects, public web data, platform scraping, sensitive domains, privacy risk, dataset sharing, consent, IRB, licenses, or data retention.
Use when preparing academic artifacts, reproducibility packages, artifact evaluation submissions, open science materials, code/data release, model cards, dataset cards, or replication bundles.
Verify in-text citations, references, author identities, year disambiguation, DOIs, and specified formatting rules; does not assess whether sources support claims. Use when the user asks for "check citation format", "verify in-text citations and references", "check authors with the same surname", "conduct a citation audit", or requests the rw-citation-audit workflow. Runs without a private local workspace or preset research-lab; use user-provided material and bundled public-source methods.
Extract academic tone rules from the corpus provided by the user in this round or the current manuscript, and make minimal adjustments. Use when the user asks for "Extract my PhD tone", "Check if the paper sounds like me", "Preserve author fingerprint", or requests the rw-phd-tone workflow. Runs without a private local workspace or preset research-lab; use user-provided material and bundled public-source methods.
Use this skill to access Reddit's full data archive via reddapi.dev API. Features semantic search, subreddit discovery, and real-time trend analysis. Perfect for market research, competitive analysis, and niche opportunity discovery.
Expert guidance for distributed training with DeepSpeed - ZeRO optimization stages, pipeline parallelism, FP16/BF16/FP8, 1-bit Adam, sparse attention
NVIDIA's runtime safety framework for LLM applications. Features jailbreak detection, input/output validation, fact-checking, hallucination detection, PII filtering, toxicity detection. Uses Colang 2.0 DSL for programmable rails. Production-ready, runs on T4 GPU.
Generate statistical analysis code with 4-round review. Select appropriate statistical tests, interpret results, and produce analysis reports with p-values, effect sizes, and confidence intervals. Use when analyzing experimental data for a paper.