product-analysis

Find, build and launch A/B tests on a Noibu-monitored store: ranked test ideas from 30 days of Noibu data and the live site, flag-gated variation code, a pull request or Shopify preview theme, automated QA, and a started test. Use for "what should we A/B test", CRO or experiment ideas, building a specific test, implementing a test already found, or taking a Noibu draft test live. Requires the Noibu MCP connector.

Audit a website for WCAG 2.1 AA accessibility violations using automated scanning and Noibu traffic data, and optionally draft the fixes. Use when you want to know what's inaccessible across the store, check WCAG or ADA compliance, find contrast or heading errors, audit for screen reader support, generate an accessibility report, or fix accessibility violations with pull requests or tickets.

Builds a rich business-context Markdown file (a CLAUDE.md / context.md) so Noibu skills and Claude stop hallucinating and reason with real context about the business. Use this skill whenever someone wants to create, set up, or improve their business context file, onboard a new store to Noibu, 'build my MD file', 'make a context file', 'set up my CLAUDE.md', capture company background for the AI, or asks why Claude keeps getting their business wrong. Also trigger on looser phrasings like 'help Claude understand my business', 'document my store for the AI', 'what should I tell Claude about my company', or when a user is starting with Noibu and has no context file yet. Even if the user just says 'build the MD file' or 'set up context', use this skill. It does deep web research on the business (always via a subagent), interviews the user about goals, money leaks, deployment process, and team setup, writes a clean Markdown file, then persists the context to the Noibu backend for the domain.

Analyze checkout performance and health using Noibu data. Use when you want to know where shoppers drop off in the checkout funnel, what payment or delivery methods customers are using, why completion rates are low, what priority errors are hurting checkout, or how cart and order values benchmark.

  Required reference for calling `noibu_*` tools — field and measure semantics,   query constraints (the `orderBy` requirement, row caps,   unique-measure rule), the call-metadata argument convention (`rationale`, `skill`), and domain   resolution. Also covers parsing of pasted `console.noibu.com` URLs.   Loads alongside topical Noibu skills   (e.g., `checkout-analysis`, `tech-diagnosis`), not instead of them — they   own routing for their respective domains.

Analyze segment and traffic performance using Noibu data. Use when you want to know which channels, devices, or countries convert best, how mobile compares to desktop, where your traffic is coming from, which segments are underperforming, or where to find your best and worst converting customer segments.

An instant, at-a-glance read on your store's health from the last 24 hours — with the option to save it as a live Cowork dashboard artifact or schedule a recurring report. Use to check on how the overall store is doing, save a live dashboard, or schedule a recurring summary.

Diagnose technical issues and Core Web Vital performance problems using Noibu data, and recommend a fix. Use when you want to know what's broken across the store, why an error is happening, what's causing errors, the root cause of poor page performance, how to fix a technical issue, why an error rate spiked, or where errors are concentrated.

Performs deep Root Cause Analysis (RCA) on NVIDIA TAO Visual ChangeNet classification experiments with image-evidence-driven investigation. Use when analyzing ChangeNet model failures, investigating poor recall / FAR / PASS-NO_PASS metrics, auditing visual inspection pipeline quality, or running an RCA report for an AOI defect-detection model. Trigger phrases include "RCA on my ChangeNet model", "why is my AOI model failing", "audit ChangeNet predictions", "investigate FAR regressions", "root cause analysis on visual-changenet".

Routes the weakest VCN samples (output of `tao-analyze-gaps-visual-changenet`) into per-augmentation-module subsets based on each module's label eligibility. Use when the user asks to "route VCN gap samples", "split AOI gaps for k-NN mining and AnomalyGen", or prepare the immediate next step after DEFT gap analysis in a VCN AOI SDA iteration.

Kumo-Anomaly — diffusion-based anomaly detection and fine-tuning for multivariate time series. Use when the user asks to "fine-tune Kumo-Anomaly", "run Kumo-Anomaly inference", "detect anomalies with diffusion", "time series anomaly detection", "finetune Kumo-Anomaly", "use perform_anomaly_analysis_with_diffusion", "automl Kumo-Anomaly", "hyperparameter search Kumo-Anomaly", "hyperparameter optimization" or mentions "curriculum_medium.yaml", "final_model.pth", "kumo-anomaly", "sdk/anomaly_analysis", or "TSDiffuser_Generic".

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.

Performs gap analysis on NVIDIA TAO VCN Classify (Visual Component Net) experiments by invoking the pinned TAO data-services container directly via `docker run … gap_analysis vcn_aoi …` — picks the optimal decision threshold, ranks per-sample weakness, and emits a top-K weakest parquet expanded per-lighting for downstream augmentation. Use when analyzing VCN classification failures, picking SDA augmentation targets, or auditing PASS/NO_PASS boundary cases.

Run TAO Data Services object-detection gap analysis from ground-truth and inference annotations. Use when an object detection workflow needs to identify weak images by comparing model predictions against ground truth using per-class recall, precision, and AP50 thresholds. Use when the user asks to "analyze OD gaps", "find weak OD images", or "run mAP gap analysis".

Brev instance operating guidance for NeMo-RL agents working in /home/ubuntu/RL with limited workspace disk, a larger /ephemeral volume, and optional /home/ubuntu/RL/.env secrets. Use when running nemo-rl-auto-research campaigns, experiments, training jobs, model or dataset downloads, shared cache-heavy commands, log-producing runs, checkpoint generation, W&B or Hugging Face authenticated workflows, or any workflow that may create large files on Brev.

NVIDIA CDF curation and retrieval. Cosmos Curator video and image pipelines (split, dedup, shard, annotate) with filtering, captioning, and embedding. Cosmos Dataset Search ingest, search, and capability qualification. TAO data mining embeddings and neighbor matching. Ships domain event catalogs, including workplace employee conduct monitoring, which requires legal review before use. Use when a request names Cosmos Curator, Cosmos Dataset Search, CDF, or their Make and CLI surfaces. Not for model training, orchestration, or unrelated dataset questions.