Instruments code so production behavior is visible and diagnosable. Use when adding logging, metrics, tracing, or alerting. Use when shipping any feature that runs in production and you need evidence it works. Use when production issues are reported but you can't tell what happened from the available data.

Use when CI is red on neilotoole/sq and you intend to act on it: an automated autofix pass, a nightly or dispatch failure, or a failing check on a pull request. Carries the triage order, what counts as proof that a fix works, and the rules that keep an attempted fix from landing as a duplicate or unmergeable PR.

Use when reviewing a pull request or a diff in the neilotoole/sq repository. Carries the repo conventions a reviewer cannot infer from the diff itself: prose and spelling rules, the split between what lint catches and what it does not, test gating, generated fixtures that must not be hand edited, and commit and PR requirements.

Guides use of the sq CLI to query SQL databases and tabular files with SLQ (sq's jq-like query language) or native SQL, manage sources, choose output formats, and run inspect, diff, and table commands. Use when the user mentions sq, SLQ, wrangling CSV/Excel/JSON/DB data, cross-source joins, or command-line data pipelines after installing sq from https://sq.io.

Use when Comet change 已完成 build 阶段，需要验证实现、处理验证失败决策或完成分支收尾。

当用户提供研究资料、项目背景、实验结果、论文草稿、PR/仓库信息或自然语言线索，希望在文献调查基础上提出科学问题、凝练可证伪假设、寻找创新点或判断研究想法价值时使用。⚠️ 不适用：用户只需要完整实验方案/分析计划（优先 research-plan）、只要写文献综述正文（优先 research-literature-review）、或只要不需要文献依据的普通头脑风暴。

Automate web browser interactions using natural language via CLI commands. Use when the user asks to browse websites, navigate web pages, extract data from websites, take screenshots, fill forms, click buttons, or interact with web applications.

Inverse-fold a backbone with SolubleMPNN — ProteinMPNN retrained on a soluble-PDB subset (Dauparas et al. 2022) — for sequences biased toward cytosolic expression and reduced aggregation. Reach for this skill when designs from vanilla ProteinMPNN are aggregating or going to inclusion bodies, when redesigning a membrane-adjacent fold for soluble expression, or when an E. coli expression screen is the next step.

Predict genome-wide functional tracks (RNA-seq, CAGE, DNase, ChIP) from DNA sequence with Borzoi. Use this skill when: (1) Scoring the regulatory effect of a variant on expression/accessibility, (2) Generating predicted coverage tracks for a locus, (3) Prioritising non-coding variants by predicted track delta.

Register a model service in the managed family — a local model server container the daemon starts/stops on demand, or a remote upstream model API (https). Read the runbook, allocate a port (local only), compose idempotent start/stop scripts (local only), register once. Load when the user wants a model service available for inference, or when list_compute shows managed endpoints.

Structure prediction using OpenFold3, an open-weights PyTorch reproduction of AlphaFold3 from the AlQuraishi Lab. Use this skill when predicting protein/nucleic-acid/ligand complex structures with an Apache-2.0-licensed AF3 reimplementation.

PyTorch-based TAO image classification. Supports a wide range of backbones (FAN, EfficientNet, ResNet, etc.) with distillation and quantization for deployment. Use when training, evaluating, distilling, quantizing, exporting, or running inference for a TAO image-classification (PyT) model. Trigger phrases include "train image classifier", "TAO classification", "ResNet/EfficientNet/FAN backbone classifier", "classification-pyt".

DINO (DETR with Improved DeNoising Anchor Boxes) for 2D object detection. Transformer-based detector with denoising training, multi-scale features, and optional distillation support. Use when training, evaluating, exporting, distilling, quantizing, or running inference for a TAO DINO detector. Trigger phrases include "train DINO", "DETR object detection", "TAO 2D detection", "DINO with distillation".

Deformable DETR for 2D object detection. Uses deformable attention for efficient multi-scale feature processing, lighter than DINO with competitive accuracy. Use when training, evaluating, exporting, quantizing, or running inference for a TAO Deformable-DETR model. Trigger phrases include "train deformable-detr", "Deformable DETR object detection", "lightweight DETR detector".

Answer what the TAO Skill Bank plugin can do by generating the response from packaged application, data, model, AutoML, and platform manifests. Use when the user asks "what can TAO Skill Bank do", "list TAO models", "which TAO workflows are available", or "what supports AutoML".

Use to summarize a recorded video via the LVS summarization microservice (HITL-gated) with a VLM fallback. Not for report generation or live RTSP captioning.

Use to deploy the vss-behavior-analytics service standalone (entrypoint, config-source, optional calibration). Not for the full warehouse deploy.

Converts cuTile Python GPU kernels (@ct.kernel) to cuTile.jl Julia equivalents. Handles kernel syntax translation, 0-indexed to 1-indexed conversion, broadcasting differences, memory layout (row-major to column-major), type system mapping, and launch API differences. Use when converting, porting, or translating cuTile Python kernels to Julia cuTile.jl, or debugging/optimizing existing Julia cuTile translations.

Recommend and customize Megatron Bridge library and benchmark recipes for a user's model, GPU count, hardware, sequence length, and pretrain/SFT/PEFT goal. Use when selecting a starting recipe, comparing library and benchmark configs, resizing parallelism for a GPU allocation, or distinguishing convergence changes, semantics-preserving execution tuning, and benchmark-only shortcuts.

Switch the active Jetson target-platform pointer to an existing profile YAML. Use before customize/build/flash to change target; not for authoring profiles — use jetson-init-target instead.