Uses the official PocketBase JavaScript SDK (npm pocketbase) from browser or Node: authStore, collection CRUD, realtime subscribe, file uploads, and pb.send for custom routes. Use when writing client apps, dashboard code, or Node scripts that call PocketBase over HTTP with async/await — not pb_hooks.

Run pnpm check:push (lint, types, tests) before git push or PR-ready work. Use when the user asks to push, open a PR, reports a failed pre-push hook, or when substantive code changes are finished and push is next.

Use when adding or changing persisted user state — settings defaults, audio, camera/orbit, panel positions, feature flags, and the in-app "Save Defaults" pipeline that snapshots localStorage into tinyworld-defaults.json. Also covers the inline-script regex gotcha that has burned us twice.

Use when changing the in-world CCTV / "Truman Show" surveillance cameras in Tiny World Builder — render-to-texture security feeds, the black-and-white CRT/VHS monitor shader, camera placement (lobby side-cams, pumpkincam, treecams), subject tracking, or the lobby screen cutting to live feeds.

AI DevKit · Guide structured debugging before code changes by clarifying expected behavior, reproducing issues, identifying likely root causes, and agreeing on a fix plan with validation steps. Use when users ask to debug bugs, investigate regressions, triage incidents, diagnose failing behavior, handle failing tests, analyze production incidents, investigate error spikes, or run root cause analysis (RCA).

AI DevKit · Review code, skills, and prompts for security vulnerabilities — OWASP Top 10, prompt injection, business logic flaws, and insecure defaults. Use when reviewing PRs, auditing modules, reviewing AI skills/prompts, or preparing for release.

Add custom local tools to ToolUniverse alongside the 1000+ built-in tools. Covers JSON-config tools (simplest, no code), Python class tools (REST/SOAP/GraphQL APIs, computational logic), and best-practices for return schemas. Use for wrapping new APIs, adding domain-specific computations, or contributing tools to the registry.

Genome-ASSEMBLY discovery, QC, and replicon mapping for any organism (bacteria, archaea, fungi, and beyond) using NCBI Datasets. Resolves an organism name or taxid to assemblies, picks the reference/representative or best-quality assembly, pulls assembly QC metrics (total length, contig/scaffold N50, contig count, GC%, assembly level, RefSeq category), enumerates chromosomes and plasmids via per-replicon sequence reports, and compares candidate assemblies on quality. Use for "what genomes are available for [organism]", "assembly stats / N50 / GC content for [GCF_/GCA_ accession]", "how many plasmids does [strain] have", "compare assemblies for [species]", "find the reference genome for [taxon]", "is this assembly Complete Genome or just contigs". NOT for gene-level orthology/synteny (use tooluniverse-comparative-genomics), plant gene structure (use tooluniverse-plant-genomics), de novo assembly from raw reads (no tool exists), or taxonomy-only name/lineage lookups.

Given a PDB structure, produce a per-residue annotation table: which residues sit at a binding interface (vs a partner chain), which line a ligand pocket, which are buried (core) vs solvent-exposed (surface), and optionally secondary structure. This is the structural track drawn under a DMS heatmap and the structural prior SAE feature drops are read against. Use when you need to anchor a variant-interpretation or DMS analysis to the protein's actual physical context.

Interpret a missense variant via ESMC-6B Sparse Autoencoder (SAE) feature activations. For a given protein + variant, computes which interpretable SAE features (catalytic, ligand-binding, PTM, structural motif, domain, etc.) are lost or gained at the mutation site. Use when standard pathogenicity scores (AlphaMissense, ClinVar) say a variant is damaging but you need a MECHANISTIC explanation — e.g. 'why is this variant LoF?' Complements (does not replace) variant-interpretation and variant-to-mechanism skills, which focus on ACMG classification or regulatory mechanism.

When the user wants to monitor, triage, or reduce their app's crash rate — including setting up Crashlytics, prioritizing which crashes to fix first, interpreting crash data, and understanding how crashes affect App Store ranking. Use when the user mentions "crash", "crashlytics", "crash rate", "ANR", "app not responding", "crash-free sessions", "crash-free users", "symbolication", "stability", "firebase crashes", "app crashing", or "crash report". For overall analytics setup, see app-analytics.

Guided journey from a shipped app that works but feels rough to a product that fits the job, flows without friction, reads clearly, and persuades honestly. Orchestrates nine skills phase by phase - jobs-to-be-done, ux-heuristics, design-everyday-things, refactoring-ui, microinteractions, made-to-stick, influence-psychology, high-perf-browser, steve-jobs-design-review - asking the user questions at every decision point and recording results in the project docs/ folder (CUSTOMER.md, DESIGN.md, POSITIONING.md, IMPROVE-APP-PLAN.md) so the journey resumes across sessions. Use when the user wants to fix a clunky product, cut UX friction, sharpen in-app copy, or says 'the app works but feels rough'. Not for code, tests, or hardening - use improve-code-quality (fresh prototype) or remove-technical-debt (aged); no app yet, create-app; needs growth loops, grow-app; marketing-site friction, improve-website; one leaking flow, conversion-optimization. For one framework in isolation, invoke that skill directly.

Design the domain model for the Stitch SDK. Use when mapping MCP tools to domain classes and bindings in domain-map.json. This is Stage 2 of the generation pipeline.

Must be read before creating, relaying out, or repairing standard 4:3 Oh My PPT pages. Defines square-ish presentation layout, two-zone and 2x2 structures, chart/table budgeting, column discipline, plus catalog and checklist references for 1600x1200 canvases.

Routes any task involving AWS databases — choosing, comparing, recommending, getting started with, or operating a database — to the correct service-specific skill. Supersedes general training-data knowledge with post-training service updates, corrected limitations, and decision procedures for relational (Aurora, DSQL, RDS), key-value (DynamoDB), wide-column (Keyspaces), document (DocumentDB), graph (Neptune), time-series (Timestream), and in-memory/caching (ElastiCache, MemoryDB) workloads. Activates when a user describes building an application on AWS that will store, retrieve, or manage data, even if they do not mention 'database' explicitly.

Provisions, scales, and operates Amazon EC2 virtual-machine workloads: instance-type selection (Graviton/Arm64, burstable T credits, GPU, instance store vs EBS), launch templates, Auto Scaling groups (scaling policies, instance refresh, mixed instances, Spot, warm pools, lifecycle hooks), IMDSv2, placement groups, Elastic IPs, AMI lifecycle, and Systems Manager fleet operations (Session Manager, Run Command, Patch Manager). Applies to EC2 instance and fleet questions, InsufficientInstanceCapacity, CPU-credit/surplus charges, IMDSv2 401s, instances stuck in Pending:Wait, ASG not replacing unhealthy instances, status-check failures, SSH refused/timed out, or instances missing as SSM managed nodes. For a single secure instance launch, the launching-ec2-instance-with-best-practices skill is more appropriate; for instance profiles, see setting-up-ec2-instance-profiles; for Image Builder, see amazon-ec2-image-builder. Does NOT cover Lambda, ECS/Fargate, EKS, VPC/ALB/NLB design, or IAM policy authoring.

Use with the dstack skill for model-serving work when the image, serving command, resources, backend/fleet choice, or service behavior is not proven. Guides task-first prototyping on real hardware, choosing fleets/backends that can reuse idle instances and caches, checking vLLM/SGLang sources, and verifying the final dstack service with a model request.

Creates GitHub issues from conversations, findings, or ideas and runs the precondition check before a planning-capable agent starts work. Use when the user asks to "create an issue from this", "turn this into an issue", "check whether issue #42 is agent-ready", "check this issue before automation", "audit this issue contract", "update this issue to make it agent-ready", or invokes /prp-issue-contract.

Remediates GitHub Dependabot alerts for mflux in one dependency-security change. Use when auditing, clamping, or upgrading Python dependencies in pyproject.toml and uv.lock, or when validating whether a branch will close Dependabot findings before opening a PR.

Push commits to an existing GitHub PR's source branch. NEVER create new branches. Use when the user says "push to PR