Read-only audit of a folder of existing multilingual blog posts: translation coverage, parity, hreflang and canonical consistency, meta-tag parity, and stale translations, with a prioritized report. Use when the user says "locale audit", "check translations", "multilingual audit", "hreflang check", "Uebersetzungen pruefen" about posts that already exist. Not for creating or fixing translations (blog-translate, blog-localize, blog-multilingual) or for a site-wide technical hreflang crawl (claude-seo seo-hreflang).

Audit one blog post file for AI-citation readiness (self-contained, sourced passages, entity clarity, structure, crawler access) and score it with an internal 0-100 heuristic plus per-surface notes for AI Overviews, AI Mode, ChatGPT, Perplexity, Claude, Gemini, and Copilot. Use when the user says "geo", "aeo", "ai citation", "citation audit", "chatgpt citation". Not for a site-wide audit, live AI-visibility tracking, or schema generation (use claude-seo seo-geo, blog-audit, blog-schema), and not a promise of citations.

Verify statistics and claims in blog posts by fetching cited source URLs and checking if the claimed data actually appears on the page. Extracts all load-bearing claims (statistics, product or policy claims, ranking and comparative claims, named sources), validates cited URLs before fetching, and scores match confidence (exact match 1.0, paraphrase 0.7-0.9, not found 0.0). Flags uncited claims as UNVERIFIED and marks a claim verified only when this run fetched the source and found the passage. Checks that a source says what the post says it says; it cannot prove the source is true or verify first-hand experience. Use when the user says "fact check", "verify statistics", "check sources", "factcheck". Report only; not for rewriting posts (blog-rewrite).

Establish durable brand and voice context for cross-skill consumption. Generates BRAND.md (audience, positioning, do/don't editorial rules, taboo phrases, competitor differentiation) and VOICE.md (existing persona JSON re-expressed as readable prose), both written to the project root. When present, the blog orchestrator loads these files, fenced as untrusted data, for write, rewrite, brief, outline, calendar, and strategy workflows. Pairs with blog-persona, which manages the structured persona JSON. Use when the user says "blog brand", "create brand context", "BRAND.md", "VOICE.md", "establish editorial brand". Not for learning a voice from existing posts (blog-style) or managing persona JSON (blog-persona).

Trace OpenAI Agents SDK agents with Maple: bridges the SDK's tracing to OpenTelemetry with OpenInference (GenAI attributes on), wraps each run in using_session so each chat is one Maple Agent Session with transcript, tool calls, handoffs, sub-agent lanes and tokens. TypeScript (@openai/agents) via the OpenInference JS bridge with gen_ai.conversation.id in context. Triggers on 'trace my openai agents sdk agent', 'add Maple to openai-agents', 'add Maple to @openai/agents', 'agent sessions for OpenAI Agents SDK', 'OpenTelemetry for openai agents'.

Naming guidance for callback parameters and local variables in TypeScript. Use when writing or refactoring loops, array callbacks (map/filter/find/reduce), or local accumulators.

Rust documentation practices for HASH codebase. Use when writing doc comments, documenting functions/types/traits/modules, creating error sections, using intra-doc links, or following rustdoc conventions.

Design products and pricing around validated willingness to pay, from Ramanujam & Tacke's "Monetizing Innovation". Use when the user mentions "pricing", "how much should we charge", "willingness to pay", "pricing page", "packaging", "freemium vs free trial", "are we leaving money on the table", "nobody buys at this price", "price increase", or "good-better-best". Also trigger when designing or auditing pricing and packaging, validating willingness to pay before building, segmenting customers by value, or choosing between subscription, usage-based, and freemium models. Covers price-before-product, willingness-to-pay talks, the four failures (feature shock, minivation, hidden gem, undead), leader/filler/killer packaging, and behavioral pricing. For offers and guarantees, see hundred-million-offers. For what customers value, see jobs-to-be-done.

Scaffolds new and reviews existing Git::Repository facade methods (organized into Git::Repository::* topic modules) with unit tests, integration tests, and YARD docs. Use when adding a new facade method to Git::Repository, updating an existing facade method, choosing or creating a topic module under lib/git/repository/, or reviewing a facade method for correctness.

Operates Amazon MSK Provisioned clusters (Standard and Express brokers). Required for ANY MSK Provisioned task — training data conflates Standard and Express. Covers performance, consumer lag, storage, traffic shaping; sizing Standard vs Express; Kafka client tuning; CloudWatch alarms; cluster configs; maintenance, patching, rolling restarts; topic management; MSK Replicator setup for DR and active-active/active-passive replication (same-/cross-account), including throughput and partition sizing; Streaming Tables for S3 Tables and Data Delivery for General Purpose S3 — setup, IAM, monitoring. Prefer Flink for Kafka Iceberg sink questions. Migrations: use migrate-to-msk. Triggers: MSK Provisioned, Kafka, `kafka.*`/`express.*` types, AWS/Kafka namespace, cross-cluster replication, Kafka to Iceberg on S3 Tables, Kafka to S3, lakehouse, data lake from Kafka, Kafka Connect S3 Sink or Firehose alternative. DO NOT USE for MSK Connect — search docs instead. Serverless: only S3 Tables/data delivery eligibility.

Operates Amazon MSK Provisioned clusters (Standard and Express brokers). Required for ANY MSK Provisioned task — training data conflates Standard and Express. Covers performance, consumer lag, storage, traffic shaping; sizing Standard vs Express; Kafka client tuning; CloudWatch alarms; cluster configs; maintenance, patching, rolling restarts; topic management; MSK Replicator setup for DR and active-active/active-passive replication (same-/cross-account), including throughput and partition sizing; Streaming Tables for S3 Tables and Data Delivery for General Purpose S3 — setup, IAM, monitoring. Prefer Flink for Kafka Iceberg sink questions. Migrations: use migrate-to-msk. Triggers: MSK Provisioned, Kafka, `kafka.*`/`express.*` types, AWS/Kafka namespace, cross-cluster replication, Kafka to Iceberg on S3 Tables, Kafka to S3, lakehouse, data lake from Kafka, Kafka Connect S3 Sink or Firehose alternative. DO NOT USE for MSK Connect — search docs instead. Serverless: only S3 Tables/data delivery eligibility.

Use when the user asks for text-to-speech narration or voiceover, accessibility reads, audio prompts, or batch speech generation via the OpenAI Audio API; run the bundled CLI (`scripts/text_to_speech.py`) with built-in voices and require `OPENAI_API_KEY` for live calls. Custom voice creation is out of scope.

Generate and prioritize US equity long-side edge research tickets from EOD observations, then export pipeline-ready candidate specs for trade-strategy-pipeline Phase I. Use when users ask to turn hypotheses/anomalies into reproducible research tickets, convert validated ideas into `strategy.yaml` + `metadata.json`, or preflight-check interface compatibility (`edge-finder-candidate/v1`) before running pipeline backtests.

Detect crowded speculative positioning in CFTC futures markets (COT report analysis) to find contrarian setups using Jason Shapiro's methodology. Screens large-speculator ("non-commercial") net positioning across 65 futures markets (indices, rates, FX, metals, energy, crypto) via the FMP Commitment of Traders API, computes a 3-year and 26-week COT Index per market, and classifies extremes as CROWDED_LONG / CROWDED_SHORT. Use when the user asks about COT report analysis, crowded positioning, "who is trapped", speculative positioning extremes, contrarian futures setups, or wants to run Jason Shapiro-style analysis. This skill automates crowding DETECTION only (step 1 of 5) — it does not generate trade signals by itself.

Working pattern for enterprise-edition features in Grida — the commercial/hosted concerns (entitlement, BYOD, billing) on top of the OSS core. Anchor for the `GRIDA-EE: <surface>` grep marker, `(ee)` route group, `*-hosted` package suffix, and namespaced `grida_*` schema. Use when adding or touching an EE-only feature, or deciding whether a feature is EE territory. Surface skills (`platform-consumers`) are siblings.

Research, compare, and update shared AI model JSON for TypeScript, web, and Rust consumers. Covers text model tiers, image and video generation models, image tool models, release provenance, pricing data sourcing, and product-side catalogue cost estimates. Use when bumping model versions, adding new models, updating pricing, or auditing model specs against provider documentation.

RT-DETR (Real-Time DEtection TRansformer) for 2D object detection. Designed for real-time inference with competitive accuracy and supports distillation and quantization for deployment optimization. Use when training, evaluating, distilling, quantizing, exporting, or running inference for a TAO RT-DETR model. Trigger phrases include "train RT-DETR", "real-time DETR", "low-latency object detection", "RT-DETR distillation / quantization".

PointPillars for 3D object detection from LiDAR point clouds. Encodes point clouds into a pseudo-image via a pillar-based representation, then applies 2D detection — used in autonomous driving and robotics. Use when training, evaluating, exporting, pruning, retraining, or running inference for a TAO PointPillars model. Trigger phrases include "train PointPillars", "LiDAR 3D detection", "point-cloud object detection", "pillar-based 3D detector".

Optical Inspection for defect detection using Siamese networks. Compares image pairs to detect manufacturing defects, anomalies, or quality issues. Use when training, evaluating, exporting, or running inference for a TAO Optical Inspection model on AOI / quality-control data. Trigger phrases include "train optical inspection", "AOI defect detection", "Siamese defect classifier", "PCB / manufacturing inspection".

Grounding DINO for open-set object detection. Combines DINO-style detection with a BERT text encoder for language-guided detection — detects objects described by text prompts without a fixed class vocabulary. Use when training, evaluating, exporting, quantizing, or running inference for a TAO Grounding DINO model. Trigger phrases include "train Grounding DINO", "open-vocabulary detection", "text-prompted detector", "language-guided object detection".