Skills — le catalogue de compétences de vos agents

Live world-data query engine. Queries real-time external APIs for factual ground truth — no synthesis, no opinion, no research. Routes requests through a registry of ~55 registered sources covering US government data, scholarly literature, weather and hazards, geocoding, finance and macro indicators, court records, nutrition, news events, property records, land due-diligence, academic papers, satellite imagery, and product manuals (3M+). Do not use for web research (use Sift), entity investigations (use Scout), or pattern analysis over historical signals.

Art-direction engine for text-to-image generation. Separates Style from Content to keep aesthetic identity consistent across an image series. Use when generating, rendering, or illustrating images (especially across a series), capturing the style of a reference image for reuse, or producing an image in a named/saved style. Not for captioning, image analysis for decisions (use ocas-look), or photo editing.

Safely audits and reclaims VPS/Linux disk space, investigates root-filesystem growth, and enforces backup retention when disk usage is high or maintenance is requested; use keywords disk cleanup, disk full, disk usage, snapshots, backups, stale repos, or disk spike. NOT for database maintenance beyond read-only analysis, logrotate configuration, or real-time monitoring.

Nightly journal curator. Batch-processes OCAS skill journals via relevance classification and writes curated content to journal files for the configured memory provider to ingest. Classifies each journal for filing as a verbatim journal note, structured entity/relationship data, or skip. Features re-emergence detection, two-pass stale handling, change magnitude gates, hibernation protection, and incremental cursor-based resumption. NOT for real-time memory filing, skill evaluation, behavioral pattern detection, or entity identity resolution.

Shared Cosmos3 frontend that explicitly routes Cosmos Framework and Cosmos-RL, validates runtime model/video-dataset/SLURM inputs, consumes an SQSH or packaged backend image, optionally plans explicit clean source builds, prepares checkpoints, validates the first update in-process, and returns token-weighted losses and task-aware accuracy.

Remote SLURM GPU cluster execution over SSH with sbatch/srun, Pyxis/Enroot containers, and Lustre-backed results. Use when running TAO training/eval/inference jobs on an on-prem or DGX SLURM cluster. Trigger phrases include "run on SLURM", "submit sbatch", "DGX SLURM cluster", "Pyxis/Enroot container", "Lustre dataset".

Extract false-positive and false-negative gaps from VLM binary-classification-question (BCQ, yes/no) predictions. Use when the user asks to "analyze VLM BCQ gaps", "extract VLM false positives and false negatives", or identify failure cases from a predictions JSON for DEFT root-cause analysis on a binary-classification VLM workflow.

Run the canonical NVIDIA AOI three-phase training pipeline — Phase 1 AutoML baseline (HPO), Phase 2 DEFT loop (RCA → SDG → mining → plain-train retrain), Phase 3 AutoML refinement on the DEFT-augmented dataset. Use when the user asks to "run the AOI workflow", "fine-tune my PCB AOI model end-to-end", "improve my AOI ChangeNet model", or "AOI workflow with AutoML" request — route here instead of tao-run-deft-aoi directly unless the user explicitly asks for the DEFT loop ONLY (e.g. "run JUST the DEFT loop", "skip AutoML, only DEFT"). Also handles the same three-phase pattern for non-AOI DEFT app

Standard single-step train/eval/export workflow for any TAO model. Use when training a TAO model on a dataset without iterative data augmentation, AutoML, or DEFT loops. Trigger phrases include "single train run", "train then evaluate then export", "plain TAO training", "normal training", "no AutoML", "skip the loop". Routes through the per-model SKILL.md for action specifics and through `tao-launch-workflow` for platform/credentials/dataset intake.

Run `tao-daft convert` to convert NVIDIA TAO DAFT datasets between supported formats. Do not use for non-DAFT data. Use when the user asks to convert a DAFT dataset, change DAFT format, change a TAO dataset format, or run `tao-daft convert`.

Run `tao-daft validate` to check NVIDIA TAO DAFT datasets for structure, schema, and cross-reference errors. Do not use for non-DAFT formats. Use when the user asks to validate a DAFT dataset, check DAFT schema, validate a TAO dataset format, or run `tao-daft validate`.

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".

提供实时运维监控看板免费版，支持会话查看、定时任务监控、网关健康检查及基础安全扫描。

Music Toolkit Free offers local MIDI generation, audio processing, chord progression, and basic music theory analysis for personal music creation and learning.

CSV数据分析免费版，提供快速统计与基础筛选，零外部依赖，适合轻量数据探索。Use when 需要数据分析、报表生成、统计洞察、数据可视化时使用。不适用于实时流数据处理。适用于独立开发者、企业团队和自动化工作流场景。支持中文交互，无需复杂配置即开即用。输出结果可直接使用，减少二次加工成本。提供结构化输出和错误处理机制。

仪表盘构建工具免费版是一款面向个人开发者的本地仪表盘产出方案，兼容从用户指定的数据源产出交互式静态 HTML 仪表盘，配套数据抓取脚本与基础可视化. 适用于需要dashboard builder相关能力的开发场景,包含结构化的工作流程和可复用的模板,帮助用户快速完成任务并保持代码质量. 适用于需要dashboard builder相关能力的开发场景,包含结构化的工作流程和配置指引.

数据工具箱免费版是一款面向个人开发者的数据全生命周期处置工具，覆盖数据提取、清洗变换、探索性剖析与基础可视化四大核心环节。Use when. 适用于需要data toolkit相关能力的开发场景,包含结构化的工作流程和可复用的模板,帮助用户快速完成任务并保持代码质量. 适用于需要data toolkit相关能力的开发场景,包含结构化的工作流程和配置指引.

数据格式转换器免费版,提供基础的数据格式互转能力.核心功能包括CSV转JSON(csv.DictReader方案)、JSON转YAML、YAML转JSON等常见格式互转,支持单文件转换与中文输出.适用于配置文件迁移、数据接口适配与跨系统数据交换场景,帮助开发者快速完成数据格式转换任务并保持数据完整性. 功能涵盖: format。

提供多源调研支持，包括范围界定、多向量搜索、五维来源评估及迭代深化，助力系统化深度研究（免费版）。

Analyzes acoustic features (frequency, duration, pitch, intensity) of livestock and poultry vocalizations to detect abnormal sounds such as coughing, wheezing, painful screams and hoarse calls, and outputs respiratory health risk hints. | 通过叫声分析识别畜禽呼吸道疾病等健康问题。

整合4个互补skill的营销组合包，覆盖Research等领域。组合内各skill协同工作，提供从数据输入到结果输出的端到端处理能力。针对信息检索效率低来源分散、brave-search: 存在易出错问题 (风险)等痛点提供解决方案。组合购买可节省20.0%费用，相比单独购买更经济。包含4个经质量审核的skill，整体评分81.5分。

辅助全流程撰写高质量、有深度且有据可查的研究长文，涵盖选题调研、结构设计、分段写作、标题优化及全文校审。

跨租户数据流监控管理器,数据流图+异常流告警+三级阈值+Prometheus 4项安全指标+每日泄露扫描。触发:数据流监控/泄露扫描/跨租户检测/data-flow/leakage-scan

Analyze Alibaba Cloud block storage (EBS) disk performance and fleet composition. Use it to locate performance bottlenecks (saturated IOPS or bandwidth), compare disks, instances, or availability zones to decide where to tune or resize, review disk count and capacity by category / region / billing type to inform capacity and cost decisions, check disk event history, and reach the right console dashboard for deeper drill-down. Triggers: "EBS monitoring", "disk metrics", "cloud disk performance", "IOPS analysis", "BPS analysis", "disk monitoring data", "export disk monitoring data", "export moni