ai_skills

Continue iterative work from the last ACCEPTED result plus only the new requested change, instead of replaying the conversation. Use during long multi-turn editing of a document, config, or code file; when the user says "keep going from what we had", asks for a follow-up change to something already approved, or is near plan/token limits in an iterative session. Saves one integrity-checked accepted result, builds a minimal continuation prompt, and never carries rejected drafts or transcript history forward.

Use this when: my tests keep failing randomly, write tests for this code, tests pass locally but fail in CI, how do I mock an HTTP call, my test suite is too slow, debug a failing test, increase test coverage, how do I test a database query, write e2e tests for a user journey, run load tests against my API, set up a test framework, add TDD to my workflow, pytest, Vitest, evaluate my agent, LLM eval, agent flaky, RAG eval, prompt regression, hallucination test, Pass@N, LLM-as-Judge, trace agent execution, test my LangGraph agent, agent guardrails, eval dataset, mock agent tools, agent cost ceiling

Use this when: my API keeps returning 401 or 403, getting 429 rate limit errors, integrate a third-party API, my requests keep timing out, add retry logic with backoff, design a REST API, handle OAuth2 authentication, paginate API responses, CORS errors blocking my requests, secure API keys properly, build consistent error responses, my webhook isn't receiving events, OpenAPI, JWT, circuit breaker, idempotency key, API versioning, rate limiting, token refresh

Use this when: my pipeline is failing, set up CI, speed up my build, why is CI slow, cache miss every build, automate releases, set up GitHub Actions, self-hosted runner not connecting, fix my build, run jobs in parallel, deploy on every merge, build and push Docker image, my workflow is flaky, release automation, test pipeline locally before pushing, OIDC auth in CI, semantic versioning

Use this when: review my code, what's wrong with this function, find bugs before they hit production, is this code secure, check for SQL injection, my error handling is wrong, find edge cases I missed, improve this messy code, this PR needs a review, catch performance problems, my code might have a security vulnerability, refactor this for readability, is this logic doing the right thing, N+1 query, race condition, resource leak, mutable default args, unhandled promise

Use this when: build a data pipeline, my pipeline is not idempotent, clean messy data, convert CSV to Parquet, my data has duplicates, validate schema at ingestion, pipeline fails on re-run, process files larger than memory, schedule a recurring job, upstream schema changed and broke my pipeline, migrate data between systems, query Parquet without loading it, deduplicate records, batch vs stream processing, DuckDB for analytics, choose an orchestrator, slow pandas pipeline

Use this when: get code onto my server, deploy to homelab, my deploy broke the service, set up automated deployment, zero-downtime deploy, how do I roll back, update Docker Compose services, health check failing after deploy, secrets in my deploy script, blue-green deployment, my service didn't come back up, automate deploys on git push, deploy without downtime, rsync to remote host, ZFS snapshot before deploy, Dockge stack deploy

Use this when: my UI looks amateurish, make this look better, fix the layout alignment, my design feels inconsistent, add dark mode, my buttons aren't keyboard accessible, build a landing page, my mobile layout is broken, enforce a design system, my animations are too aggressive, clean up my component styles, my UI has no visual hierarchy, add proper focus indicators, Tailwind, Shadcn, React, Next.js, accessibility audit, design tokens, WCAG, responsive, component architecture, Storybook

Guide complete software development lifecycle from requirements through production. Tech stack selection, ADR templates, branching strategy, testing pyramid, CI/CD, release management, observability setup, and architectural onboarding for new codebases.

Expert GitHub workflow automation. Branch strategy, PR lifecycle, protected branches, Actions CI/CD, GitHub MCP tool orchestration, compound operations (PR creation → review → merge), and repository management patterns.

Use this when: draft an email for me, triage my inbox, schedule a meeting and check availability, create a calendar event with an agenda, add rows to a spreadsheet, organize my day, automate a weekly email report, summarize my unread emails, find a free slot for a call, my inbox is overwhelming, send a meeting invite with context, append data to a sheet automatically, Gmail, Google Sheets, compound Workspace operations, daily briefing, report sync

Universal homelab infrastructure specialist. Docker self-hosting, TrueNAS ZFS operations, Proxmox VM/LXC management, K3s Kubernetes with FluxCD, 3-2-1 backup strategy, blue-green deployments, internal DNS (Pi-hole/Technitium), VLAN design, Home Assistant + MQTT, and cross-platform migration planning.

Manage infrastructure declaratively with Terraform and Ansible. Provision cloud and homelab resources, enforce idempotent configuration, manage secrets with Vault/SOPS, and maintain state safely across environments.

Use this when: run a model locally, which model fits my GPU, my inference is too slow, serve LLM in production, how to quantize a model, too many concurrent users, self-host an AI, set up Ollama, vLLM vs Ollama, route between multiple models, VRAM out of memory, model won't load, tokens per second too low, pick a quantization format, LiteLLM gateway setup, what model fits 8GB VRAM, OpenAI-compatible local endpoint

Use this when: build an MCP server, expose my API as a Claude tool, connect a service to Claude, my MCP tool is not showing up, write a custom Claude tool, MCP server not connecting, debug MCP connection, set up FastMCP, add auth to an MCP tool, my tool returns wrong format, register in Claude Desktop, tool is not being called, stdio vs SSE transport, wrap a REST API for Claude, configure .mcp.json, MCP resource vs tool, MCP prompt vs tool, Docker MCP container, multi-server setup, path traversal in file tool

Meta-agent orchestrator for complex multi-step projects. Routes tasks to specialized agents via DAG decomposition, resolves conflicting agent outputs with evidence-based arbitration, provides cost-aware delegation, and includes RF/ham radio calculation reference (antenna, link budget, coax, SWR, AREDN).

Build personal and team productivity systems. Email triage and drafting, calendar management, spreadsheet automation, Zapier/n8n workflow design, note-taking pipelines, and compound operations that reduce manual overhead.

Use this when: set up a homelab Kubernetes cluster, create a VM from a template, my K3s node won't join, set up GPU passthrough, automate VM provisioning with cloud-init, my IOMMU groups are wrong, back up VMs automatically, install K3s on a new node, deploy persistent workloads in Kubernetes, manage LXC containers, rebuild infrastructure from code, set up GitOps for my cluster, Proxmox, K3s, FluxCD, ArgoCD, Helm, Longhorn, PBS, Terraform proxmox, VLAN design, Ceph vs NFS, MetalLB, NVIDIA passthrough, ACS override

Use this when: is this code safe, security audit, fix a vulnerability, hardcoded secrets in code, my secrets got leaked, rotate leaked credentials, SQL injection, broken authentication, how do I hash passwords, add MFA, my container is running as root, dependency has a CVE, OWASP compliance, missing auth check, is my API secure, certificate expired, incident response, supply chain attack, vulnerability assessment, container hardening, SBOM, cosign, Trivy scan, secrets scanning, trufflehog, gitleaks, CIS benchmark, SSH hardening, firewall rules, TLS setup, distroless image

Connect and orchestrate services using n8n, Redis Streams, MQTT, webhooks, and REST. Build automation workflows, event-driven pipelines, service health monitoring with Uptime Kuma, and Traefik reverse proxy routing.

Ingest, fuse, and analyze RF/SIGINT and OSINT data streams. Build ingestion workers, normalize feeds into PostGIS/pgvector schemas, stream events through Redis Streams, and query fused intelligence. Sources: APRS, ADS-B, AIS, GDELT, RSS threat intel, SDR.

Universal software architect consolidating full SDLC, code review, database architecture, repo health, ADR management, tech debt scoring, schema design, and codebase onboarding. From requirements through production maintenance.

Comprehensive testing methodology covering strategy, frameworks, patterns, and debugging. Unit, integration, E2E, load, contract, property-based, and mutation testing. Python (pytest) and TypeScript (Vitest/Playwright/k6). Fix flaky tests, design test strategies, set up CI integration.

Use this when: fine-tune my model, my training loss won't converge, out of memory during training, LoRA vs full fine-tune, my GPU is too small, set up distributed training, model is overfitting, pick a learning rate, prepare a training dataset, QLoRA on consumer GPU, instruction tuning, run training on multiple GPUs, my val loss is diverging, pick batch size, configure DeepSpeed, what rank for LoRA, how many training epochs

Use this when: set up an SMB or NFS share, my ZFS pool shows errors, automate dataset snapshots, replicate data to another NAS, fix dataset permissions for Docker containers, my share is not accessible, migrate TrueNAS CORE to SCALE, tune storage for media or databases, add a cloud backup destination, my disk is failing, expand or replace a drive in the pool, set up offsite replication, TrueNAS, ZFS, ix-applications, SCALE 24.10, docker on TrueNAS, pool health check, dataset record size, NAS API, container bind mount permissions

Use this when: my site is slow, why is my Lighthouse score low, my LCP is too high, my page keeps shifting layout, fix Core Web Vitals, make my site accessible, keyboard navigation is broken, screen readers can't use my page, reduce bundle size, fix color contrast issues, my page fails WCAG, images are hurting my load time, why is my page janky on scroll, Lighthouse, axe, INP too slow, CLS layout shift, TTFB slow, ARIA wrong, focus indicator missing, skip link, alt text, font loading flash

Expert guidance for Android OBD2 / ELM327 development — protocol initialization, PID polling, vehicle data parsing, adapter quirks, and Android Auto integration. Use whenever the user works with OBD2 data, ELM327 commands, AT command sequences, PID requests, DTC reading, vehicle protocol detection, ATSP0/ATZ/ATE0/ATL0/ATH0/ATAT1, bus init failures, adapter init timeouts, classic OBD socket connect, OBDLink MX+/CX, Veepeak OBDCheck BLE+, generic ELM327 adapters, NO DATA, UNABLE TO CONNECT, BUS INIT, SEARCHING, CAN ERROR, STOPPED, ECU polling, race telemetry from OBD, Android Auto OBD display, or Car App Library integration with live vehicle data. Trigger when OBD appears connected but returns no valid data, or when BLE-vs-classic path choice changes behavior. Do not use for generic Android Bluetooth permissions or pairing-only issues; use android-bluetooth for those.

Use this when: scaffold a local AI orchestration stack, set up LangGraph supervisor, connect Ollama to OpenCode, configure MCP servers for agent tools, wire OpenCode as the developer interface for a multi-agent system, set up A2A protocol between LangGraph and OpenCode, configure VS Code for agent development, orchestration stack on TrueNAS / Docker Compose homelab, choose between AG2 and CrewAI sub-frameworks, set up agent harness with PreToolUse hooks, define SLIs for an AI orchestration stack, design generator-evaluator pattern for agent output, local inference routing with Qwen3 or Mistral Small.

AI Innovation Radar — strategic AI innovation scanning and advising system. Auto-trigger on ANY of these cues: 'drop' or pasting Perplexity findings/article batches for evaluation, 'briefing on [project]', 'working on [project] today', 'survey [project]', 'horizon update', 'radar', or any reference to evaluating AI tools/innovations against active projects. The five operating modes are Drop (evaluate a Perplexity batch), Briefing (deep project review), Build (focus mode for a working session), Survey (landscape scan), and Horizon (macro AI trends).

Use this when: upgrade AGP, migrate Android Gradle Plugin, AGP version upgrade, AGP 9 migration, update build.gradle AGP version, Android build plugin upgrade, gradle plugin compatibility. Upgrades, or migrates, an Android project to use Android Gradle Plugin (AGP) version 9. Do not use this skill for migrating Kotlin Multiplatform (KMP) projects.

Android BLE hardware integration using the Nordic Semiconductor BLE library (no.nordicsemi.android:ble). Covers UART/GATT service discovery, callbackFlow-based device scanning, connection lifecycle management, multi-device type detection, automatic reconnection, and byte-level protocol parsing. Use this when: integrating a BLE peripheral, implementing a GATT profile, scanning for devices by UUID or name pattern, handling BLE permissions on Android 12+, parsing binary telemetry packets, building a BleManager subclass, handling disconnect/reconnect loops, integrating GPS loggers, sensor hardware, OBD-II adapters, or any custom BLE peripheral. Do not use for general Bluetooth Classic/SPP/RFCOMM pairing or connection management — use android-bluetooth instead.

Use this when: android create project, android sdk install packages, android sdk list targets, android emulator create AVD, android emulator start, android run APK on device, android build tools, adb commands, Android project scaffolding from template, AVD management, deploy Android app from CLI, android environment info. Orchestrates Android development tasks including project creation, deployment, SDK management, and environment diagnostics using the android command-line tool.

Use this when: migrate XML layout to Compose, convert XML View to Jetpack Compose, replace XML UI with Compose, interop XML and Compose, ComposeView in XML, AndroidView in Compose. Provides a structured workflow for migrating an Android XML View to Jetpack Compose. This skill details the step-by-step process, from planning and dependency setup, to theming and layout migration, validation and XML cleanup. Use this skill when you need to migrate an XML View to Jetpack Compose in an Android project. It solves the problem of converting the UI of a legacy XML View into modern, declarative Compose components while maintaining interoperability.

Use this when: audit my business model for AI risk, where is my value at risk from AI, which parts of my business does AI threaten, arbitrage gap analysis, how durable is my competitive advantage, AI compression of my market, is my business model AI-proof, what gaps am I depending on, competitive moat analysis, where AI will eat my margin, assess my business against AI disruption, what happens to my business when AI improves, am I on the right side of AI disruption, strategic AI risk assessment

Use this when: I need current information from the web, fact-check this against live sources, find the latest news on, what changed recently in, verify this claim with citations, competitive analysis with sources, find documentation for a specific version, what is the current state of, I need links to primary sources. For deep offline research and threat intelligence triangulation, use research-analyst instead.

Use this when: set up Grafana dashboards, configure Prometheus scraping, write recording rules, configure alerting, monitoring is broken, dashboards are wrong, cardinality is exploding, set up SLOs, configure Alertmanager, GPU monitoring, AI training metrics, model inference monitoring, LiteLLM metrics, vLLM metrics, Ollama metrics, Qdrant health, vector DB monitoring, DCGM exporter, dashboards as code, provision Grafana without clicking, Grafana API, Grizzly, Grafonnet, Terraform grafana provider, kube-prometheus-stack, multi-burn-rate alerts, MWMBR, RED method, USE method, node exporter, cadvisor, blackbox exporter, Prometheus federation, remote write, high availability, Loki, Tempo, Mimir, AI workflow monitoring, training job metrics, inference monitoring, model serving SLO, drift detection, pushgateway

Develop native iOS applications with Swift/SwiftUI. Masters iOS 18, SwiftUI, UIKit integration, Core Data, networking, and App Store optimization.

Use this when: research this topic with source triangulation, fact-check this claim across multiple sources, I need citations with confidence scoring, verify this information from primary sources, competitive analysis with evidence, what does the latest research say, find threat actors behind, investigate this IOC, threat intelligence report, triangulate this finding, market research with sourcing, track adversary campaigns. For live web search and current news, use deep-research-pro instead.

Use this when: my alerts are too noisy, set up monitoring, service is down, alert fatigue, define SLOs, write a runbook, postmortem template, why is my service slow, error rate is spiking, set up observability, latency percentiles, disk is filling up, incident response, on-call escalation, trace a slow request, capacity planning, Prometheus alerting, my dashboard shows nothing useful

Use this when: when should I raise my next round, timing my fundraise around the IPO window, M&A timing around AI mega-IPOs, capital market timing strategy, when is the best time to fundraise, hiring AI talent before the IPO window, strategic decisions around upcoming AI IPOs, sequence my business decisions, capital absorption window, IPO cycle planning, lock-up expiration strategy, when to raise vs wait, avoid the IPO capital crunch, strategic timing for CEOs, board strategy around mega-IPOs