Claude Skills by aicodedecode

Ultra-low-latency inference on Groq's LPU hardware — when tokens-per-second is the product requirement.

Groundwater flow and transport modeling — aquifer characterization, MODFLOW workflows, calibration, and capture-zone analysis.

Run growth experiments — acquisition loops, activation, retention, and referral systems driven by rapid testing.

gRPC guidance — protobuf design, service patterns, streaming, deadlines, load balancing, and debugging.

Build habits with your AI assistant as coach: start tiny, stack habits onto existing routines, do quick daily check-ins, repair streaks with the never-miss-twice rule, and review weekly. Use when starting exercise, reading, meditation, or any daily routine.

Design habit systems that stick: habit selection, tracking methods, streaks, environment design, and recovery from slips. Use when building new routines or fixing ones that keep failing.

Generate and verify cryptographic hashes with algorithm selection, salting, and integrity workflows.

Idiomatic Haskell: pure functional design, type-driven modeling, monad stacks, and principled effects. Use when writing, reviewing, or structuring Haskell programs.

Build retrieval-grounded agents with Haystack — pipelines, components, and agent patterns for search-heavy workflows.

Helm guidance — chart structure, templating, values management, releases, hooks, and chart best practices.

Run your household like a calm system: capture every home task into a master list, pick a weekly room/zone focus, track maintenance schedules (filters, servicing, seasonal jobs), and split chores fairly with family or roommates. Use when the house feels chaotic, chores are unevenly split, or maintenance keeps getting forgotten.

Using managed crawl/scrape/extract APIs — when to outsource scraping and how to evaluate providers — vendor-neutral patterns.

Professional HTML: semantic markup, accessibility, forms, SEO fundamentals, and document structure. Use when writing or reviewing HTML.

Converting HTML/CSS to PDF — print stylesheets, pagination, and renderer choice — use when generating PDFs from web content.

Run models via Hugging Face Inference — serverless endpoints, dedicated deployments, and the Hub ecosystem.

Hugging Face guidance — Hub, Transformers, datasets, fine-tuning, inference endpoints, and Spaces.

Make AI-generated text sound human: de-genericizing drafts, restoring voice, and editing for natural flow. Use when AI-assisted writing sounds robotic, verbose, or obviously generated.

Combine dense and sparse retrieval — BM25 + vectors, fusion algorithms, weighting, and tuning. Use when neither keyword nor semantic search alone is good enough.

Catchment hydrology modeling — water balance, rainfall-runoff models, calibration, and flood estimation.

Affordable GPU cloud and inference on Hyperbolic — decentralized compute with competitive pricing.

Build identity and access management — lifecycle, privileged access, governance, and continuous access review.

Design clear, consistent icon sets with grid systems, stroke rules, and production-ready exports.

Deploy and tune intrusion detection/prevention — sensor placement, signature and behavioral detection, and alert quality.

Create custom illustrations with consistent style, composition, and narrative for brands and products.

Handle professional image tasks with format conversion, resizing, compression, and metadata workflows.

Retouch images to professional standards with frequency separation, dodge and burn, and natural results.

Design in-app messages — tooltips, modals, banners, and onboarding flows that guide without annoying.

Tame email overload with Muse as your coach: the 4D triage method (delete/delegate/respond/defer/file), batching, reply templates, unsubscribing, and a daily 15-minute routine.

Leading incident response — roles, communication, and decision-making under pressure — use when running or improving on-call response.

Communicate during incidents — stakeholder updates, severity frameworks, war-room comms, and post-incident reviews.

Navigate breach and incident notification duties — GDPR, NIS2, sector rules, and building the report-ready capability.

Lead security incident response end to end — preparation, detection, containment, eradication, recovery, and lessons learned.

Plan influencer partnerships — finding creators, outreach, negotiation, campaign management, and ROI measurement.

Turn complex data and processes into clear, honest infographics with strong visual storytelling.

Build durable serverless workflows with Inngest: step functions, event-driven flows, retries, and scheduling. Use when background work must survive failures and run to completion.

Get structured, validated outputs from LLMs with Instructor — Pydantic schemas, retries, and reliable JSON extraction.

Design SaaS integrations — data sync strategies, embedded experiences, marketplace apps, and partner ecosystems.

Building interactive scientific visualizations — Plotly/Bokeh dashboards, linked views, and exploratory data apps.

Build with Shanghai AI Lab's InternLM models — open models strong in reasoning, code, and Chinese.

Run realistic mock interviews with structured feedback on answers, body language, and storytelling technique.

Plan and conduct research interviews — guides, probing techniques, rapport, recording, and ethical practice. Use when gathering deep qualitative insight from conversations.

Build cross-platform mobile apps with Ionic: UI components, navigation, Capacitor integration, theming, and performance. Use when building hybrid mobile apps with web tech.

Triaging error reports — prioritization, ownership, and keeping the backlog actionable — works with any error tracker.

Analyze LLM jailbreak techniques defensively — how bypasses work, detection signals, and hardening strategies. Use for safety research and building more robust guardrails. Taxonomy and defense only; no bypass instructions.

Idiomatic modern Java: project structure, streams, records, concurrency, build tooling, and testing. Use when writing, reviewing, or structuring Java applications.

Idiomatic modern JavaScript: ES2022+ features, async patterns, modules, and runtime behavior mastery. Use when writing, reviewing, or debugging JavaScript.

Jenkins guidance — pipeline-as-code, agents, shared libraries, credentials, scaling, and hardening.

Professional Jetpack Compose: declarative UI, state hoisting, side effects, navigation, and performance. Use when writing, reviewing, or structuring Compose apps.

Automate Jira via its native REST API: bulk issue operations, custom reports, and integrations without middleware. Use when Jira's UI and built-in automation can't do what you need.

Run Jira effectively: project setup, workflows, JQL, automation, and reporting for software teams. Use when configuring Jira or improving team process around it.