AI Agent Workflow Collection

Use this when: can AI agents use my service, is my product agent-ready, agentic commerce readiness, MCP server for my service, can an AI agent buy from me, agent-native infrastructure, stress test my service for AI agents, how do AI agents discover my product, programmatic API for agents, machine-readable service description, agent-to-service interaction, make my service agent accessible, audit my service for AI agent compatibility, agentic economy, how do I sell to AI agents, discovery evaluation transaction verification…

Upgrades, or migrates, an Android project to use Android Gradle…

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: build a skill, write a SKILL.md, my skill isn't triggering, my skill over-triggers, my agent ignores instructions, improve my skill description, my skill isn't routing correctly, write skill instructions, design an AI agent, test my skill routing, optimize trigger phrases, build a system prompt, create a new agent skill, how do I structure a skill, skill routing is broken, write instructions an AI follows reliably, my skill…

Use this when: design an AI system, RAG vs fine-tuning, my agent keeps looping, architect a multi-agent system, which LLM should I use, context window keeps overflowing, add guardrails to an agent, build a production AI app, reduce AI inference cost, evaluate my AI pipeline, my agent calls the wrong tools, budget tokens for context, design agent memory, when to use MCP, pick a vector store, set up LLM-as-judge evaluation, agent architecture…

Android Auto and Automotive OS app development using the Car App Library: template-based UI architecture, voice interaction, safety restrictions (distraction optimization), session and screen lifecycle, navigation templates, media playback integration, and place list/map templates. Covers both the phone-projection (Android Auto) and native (AAOS) deployment targets. Use this when: building an Android Auto app, porting a phone app to the car screen, implementing voice commands for Auto, handling driving safety restrictions, building navigation or POI features for car, or targeting Android Automotive OS natively.

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.

Expert guidance for Android Bluetooth Classic (SPP/RFCOMM) and BLE development. Use this skill whenever the user is building, debugging, or architecting Android apps that connect to Bluetooth devices — including BluetoothSocket, RFCOMM, SPP, BLE, GATT, BluetoothAdapter, BluetoothDevice, bonded devices, paired devices, classic-vs-BLE transport selection, device scan routing, ACL disconnect, disconnect detection, reconnect loops, stuck Connecting UI, stale connected state, background Bluetooth, foreground service ownership of connection state, Android 12+ Bluetooth permissions, BLUETOOTHSCAN, BLUETOOTHCONNECT, Bluetooth connection dropping, Android app shows connected when it is not, OBDLink MX+ classic pairing, Veepeak BLE+ routing, RaceBox Bluetooth lifecycle, or coroutine/state-machine bugs in Android Bluetooth connection management. Do not use for non-Android Bluetooth firmware or low-level ECU / PID parsing questions; use android-obd for ELM327 and OBD protocol work.

High-frequency real-time data display in Jetpack Compose: recomposition avoidance strategies, Canvas-based gauge and graph rendering, StateFlow-to-Compose integration, and stable data class patterns. Covers derivedStateOf, remember keys, snapshot-based reads, and performance profiling for 25Hz+ UI updates. Use this when: displaying live sensor or telemetry data in Compose, building custom Canvas gauges or live graphs, optimizing a Compose screen that recomposes too often, streaming StateFlow values to UI at high frequency, or building a dashboard with multiple independently-updating widgets.

Patterns for long-running Android foreground services used for continuous data collection — sensor streaming, BLE sessions, audio recording, or GPS logging that must survive Activity lifecycle changes. Covers Android 14+ mandatory foregroundServiceType, STARTSTICKY null-intent safety, notification channel setup, intent-based IPC, source fallback logic, and clean shutdown. Use this when: building a recording or sensor collection service, keeping BLE alive across screen rotation, handling Android 14 foregroundServiceType enforcement, wiring a Service to a Compose ViewModel, or implementing source-priority fallback.

Expert guidance for Android OBD2 / ELM327 development — protocol initialization, PID polling, vehicle data parsing, adapter quirks, transport quirks, and Android Auto integration. Use this skill whenever the user is working with OBD2 data, ELM327 commands, AT command sequences, PID requests, DTC reading, vehicle protocol detection, 0100 probe behavior, ELM327 prompt handling, ATSP0, ATZ, ATE0, ATL0, ATH0, ATAT1, vehicle bus initialization failures, adapter init timeouts, classic OBD socket connect, OBDLink MX+, OBDLink CX, Veepeak OBDCheck BLE+, generic ELM327 adapters, clone behavior, 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 for app issues where OBD appears connected but does not return valid data, where Home shows Connecting during adapter init/probe, or where BLE-vs-classic OBD path choice changes behavior. Do not use for generic Android Bluetooth permissions or pairing-only issues unless they are directly blocking OBD transport setup; use android-bluetooth for those.

On-device machine learning for Android: ML Kit (text recognition, object detection, pose estimation, face detection), TensorFlow Lite (custom model inference with GPU/NNAPI delegate), and MediaPipe Tasks API (gesture recognition, image classification, hand/face/pose landmarking). Covers model packaging, interpreter lifecycle, input/output tensor handling, Compose integration, and performance optimization for real-time inference. Use this when: running ML inference on-device without a server, integrating ML Kit APIs, loading a custom TFLite model, using MediaPipe for vision tasks, building a camera + ML pipeline, optimizing inference latency, or handling model versioning and asset delivery.

Android security fundamentals: encrypted storage (EncryptedSharedPreferences, EncryptedFile), biometric authentication (BiometricPrompt, Credential Manager), network certificate pinning, SafetyNet/Play Integrity API, KeyStore-backed key generation, and secure data handling patterns. Use this when: storing sensitive data locally, implementing biometric login, pinning TLS certificates, protecting API keys, integrating Play Integrity attestation, generating cryptographic keys in Android KeyStore, or hardening an app against reverse engineering.

High-frequency sensor data pipeline for Android: hybrid storage (Room for session metadata + CSV/binary files for sample data), StateFlow-based streaming to UI, session lifecycle management, and offline-first write patterns. Covers atomic session creation, buffered CSV writes, telemetry normalization across sources (BLE device vs phone GPS+IMU), and post-session data analysis. Use this when: ingesting 10-100Hz sensor data, designing a session storage schema, streaming live telemetry to a Compose ViewModel, building a session list/review flow, handling source switching mid-session, or implementing post-session analysis pipelines.

Android video and photo capture using CameraX: lifecycle-aware camera setup, Compose camera preview, VideoCapture use case with MediaStore output, simultaneous preview + capture, camera selector, torch control, and recording state management. Use this when: building a camera screen in Compose, implementing video recording with audio, saving video to MediaStore (gallery), capturing photos, switching front/back cameras, controlling the torch/flashlight, or managing concurrent camera use cases.

Use this when: save this page for later, archive websites before they disappear, my bookmarks are disorganized, prevent link rot, set up ArchiveBox, build a personal knowledge base, capture web pages offline, I'm losing track of articles, import my Pocket bookmarks, dedup my archive, archive RSS feeds, search my saved articles, capture JavaScript-heavy pages, organize my reading list, set up a read-later system, OCR my documents, build a second…

Use this skill to migrate legacy Android camera implementations…

Use this when: audit my career for AI risk, is my job at risk from AI, which of my skills are AI-proof, am I on the right side of AI disruption, career strategy for AI era, future-proof my career, what skills should I develop, is my role being automated, AI impact on my job, career gap analysis, where should I focus my skills, what do I do with AI time savings, am I using AI correctly in my career, career migration plan, am I riding a closing gap, how do I stay relevant, reskill for AI, career readiness…

CI/CD pipeline guide covering GitHub Actions, Lefthook hooks, Dependabot, and deployment scripts. Use when asking about workflows, automation, or deployment.

MANDATORY when designing schemas, writing migrations, creating indexes, or making architectural database decisions - enforces PostgreSQL 18 best practices including…

Use this when: set up a self-hosted service, my container keeps restarting, add HTTPS to a container, connect containers to each other, set up a reverse proxy, my volumes are not persisting, migrate a stack to a new server, expose a service to the internet, back up container data, deploy on TrueNAS, run multiple services on one host, containers can't reach each other, Docker, Traefik, Cloudflare…

Use this when: give my agent persistent memory, my agent forgets between sessions, set up vector search, choose a vector database, long-term memory for AI agents, agent can't recall past conversations, build a second brain for AI, share memory between agents, my embeddings are outdated, search past conversations semantically, add a knowledge graph, memory search is slow, store decisions permanently, Qdrant vs pgvector, deduplicate memory entries, memory grows without bound, federate memory across…

Use this when: am I a middleware wrapper, is my AI startup defensible, what's my moat against AI, positioning audit for AI era, five verticals analysis, trust context distribution taste liability, am I in the middleware trap, how do I survive when models get better, is my product AI-proof, strategic positioning for AI company, competitive moat AI, what happens to my product when foundation models improve, structural durability of my business, is my SaaS defensible against AI, where is my real value, AI era business strategy…

Use this when: start a new project, how should I structure this, what stack should I use, monolith vs microservices, set up CI/CD, testing strategy, automate releases, semantic versioning, my project has no tests, set up pre-commit hooks, feature branch strategy, scaffold a new service, when should I split into services, project is hard to onboard, document architecture decisions, release…