Skills for “ci”

…from any source connected to Windsor.ai (ads, organic social, analytics, CRM, e-commerce, finance) to warehouses, databases, spreadsheets, BI tools, or cloud storage - BigQuery, Snowflake, Redshift, PostgreSQL,…

…inputs, show the full draft, and create only after explicit…

…constraints. Use class_stratified when rare classes are…

…models, using Nano by default and Edge or Super when explicitly requested: evaluate the base model on Proxy and frozen Benchmark splits, mine real image pairs from Proxy gaps, generate AnomalyGen synthetic NG pairs,…

…and the migration jobs (discover, structure, reconcile, plus one copy job per schema of the approved plan, each passing its schema as a task parameter). Use when the user wants to set up AIDP for the…

…source. Auth is OAuth client-credentials (Service Principal client_id + secret +…

…use the resolved image; venv training requires an explicit request. Platform-agnostic — runs on any SDK (Brev, SLURM, Kubernetes, Docker). Do not use generic "keep improving" language alone to override a matching…

…a vision / VLM / LLM model end-to-end, generate a reproducible HF training pipeline, smoke-test a HuggingFace model locally before scale-up, push a fine-tuned model to the HF Hub with a model card, or emit a…

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…

Validate a user AnomalyGenNext dataset and recipe, freeze its validation set, and prepare a canonical texture fine-tuning recipe. Use when AnomalyGenNext task weights do not yet exist. Do not use for ordinary defect…

Token-efficient execution kit for repeated TAO agentic workflows. Compiles a skill into stage cards once (with a strong model), then executes them forever in fresh, small headless sessions coordinated by a bash driver —…

…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…

…against a migrated AIDP target via `infa2aidp cli reconcile`, driven by a YAML config (row_count / schema / aggregate / data checks, with ignore_columns for audit/surrogate columns that never match). Use after a…

…pass, not a correctness check (that's infa-reconcile) and not a cluster-execution check (nothing here touches a live…

…using FastFoundationStereo (FFS), the distilled bp2 commercial variant of FoundationStereo. Predicts disparity maps from stereo image pairs with ~10× lower latency than full FoundationStereo. Use when training,…

…job, when the operator runs it. Use when the user has explicitly asked for a Standard catalog, or wants the migration delivered as a runnable script rather than executed straight from the…

…conversion and REVIEW REQUIRED marker for a human decision (approve/edit/reject) rather than letting them ship…

…an approved Snowflake plan only when the user has explicitly asked for a Standard catalog. This stage copies no data - rows are copied only by the snowmig_02_copy_schema job, when the operator runs it. Use when the…

…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…

Start, query, and stop a network-specific TAO inference microservice ({network_arch}-inference-microservice) by delegating container execution to the appropriate platform skill. Handles container image resolution,…