HOUSTON executes feature tasks directly. No agents. Best for small features (1-3 tasks).

Guide for configuring Mixins in Architectury multi-loader projects

Interpret machine learning models using SHAP, LIME, feature importance, partial dependence, and attention visualization for explainability

Diagnose and stabilize ML training runs, recover from failures, and deliver validated fixes.

MLflow 3 GenAI evaluation for agent development. Use when (1) writing mlflow.genai.evaluate() code, (2) creating @scorer functions, (3) building evaluation datasets from traces, (4) using built-in scorers (Guidelines, Correctness, Safety, RetrievalGroundedness), (5) analyzing traces for latency/errors/architecture, (6) optimizing agent context/prompts/token usage, (7) debugging evaluation failures. Covers the full eval workflow: trace analysis -> dataset building -> scorer creation -> evaluation execution.

Enterprise practical workflows, context engineering strategies, JIT (Just-In-Time) retrieval optimization, real-world execution examples, debugging patterns, and moai-adk workflow mastery; activates for workflow pattern learning, context optimization, debugging issue resolution, feature implementation end-to-end, and team knowledge transfer

Use when the user needs to query, filter, reshape, extract, create, or construct JSON data — including API responses, config files, log output, or any structured data — or when helping the user write or debug JSON transformations.

Context preservation, tiered summarization, resumability

'Expert guidance on building and deploying apps to WendyOS edge devices. Use when developers mention: (1) Wendy or WendyOS, (2) wendy CLI commands, (3) wendy.json or entitlements, (4) deploying apps to edge devices, (5) remote debugging Swift on ARM64, (6) NVIDIA Jetson or Raspberry Pi apps, (7) cross-compiling Swift for ARM64.'

'Reverse engineers .NET malware using dnSpy decompiler and debugger to analyze C#/VB.NET source code, identify

Debugs native module crashes, optimizes V8 performance, configures node-gyp builds, writes N-API/node-addon-api bindings, and diagnoses libuv event loop issues in Node.js. Use when working with C++ addons, native modules, binding.gyp, node-gyp errors, segfaults, memory leaks in native code, V8 optimization/deoptimization, libuv thread pool tuning, N-API or NAN bindings, build system failures, or any Node.js internals below the JavaScript layer.

Run, rerun, debug, or interpret OpenClaw Parallels install, onboarding, gateway smoke, and upgrade checks.

Write reusable ChatCrystal task memories after substantive work completes. Use when implementation or debugging produced a durable fix, pitfall, pattern, or decision worth preserving, and when the environment can either persist it through `write_task_memory` or emit a structured memory candidate for later save.

Systematic diagnosis of bugs using the performance-profiler agent

**Version**: 1.0.0

Deployment and build automation. Always push when building. Verify branch (main) before push. Deploy to production after changes. Use for: build, deploy, push, release, ship. Monorepo: check Vercel Git integration, root directory, GitHub Actions fallback.

debug

Coordinates parallel investigation threads to simultaneously explore multiple hypotheses or root causes across different system areas. Use when debugging production incidents, slow API performance, multi-system integration failures, or complex bugs where the root cause is unclear and multiple plausible theories exist; when serial troubleshooting is too slow; or when multiple investigators can divide root-cause analysis work. Provides structured phases for problem decomposition, thread assignment, sync points with Continue/Pivot/Converge decisions, and final report synthesis.

Build, debug, and optimize CMDx tasks and workflows in Ruby. Use when creating service/command objects with CMDx, composing business logic into workflows, handling task failures and interruptions, or working with CMDx attributes, callbacks, middleware, and configuration.

4-phase systematic debugging methodology with root cause analysis and evidence-based verification. Use when debugging complex issues.

Design and maintain the Phase 01 Orchestra YAML registry, Rust core types, atomic save/load, and CLI bootstrap commands.

Create and manage Claude Code skills following Anthropic best practices. Use when creating new skills, modifying skill-rules.json, understanding trigger patterns, working with hooks, debugging skill activation, or implementing progressive disclosure. Covers skill structure, YAML frontmatter, trigger types (keywords, intent patterns, file paths, content patterns), enforcement levels (block, suggest, warn), hook mechanisms (UserPromptSubmit, PreToolUse), session tracking, and the 500-line rule.

Debugging techniques and tools for IntelliJ development. Use when debugging IDE issues or investigating failures.

This is the authoritative architecture document for any AI agent or developer working in this codebase. Read this before reading source files.

- **Skill Name**: exploit-development

Crafting effective prompts for LLMs. Use when designing prompts, improving output quality, structuring complex instructions, or debugging poor model responses.

Identifies mobile layout issues across devices: responsive breakpoint problems, text sizing issues, column stacking failures, element visibility problems, and navigation menu behavior.

'Use this skill whenever the user installs, configures, or debugs Google Analytics 4, or mentions gtag.js, GTM, conversion tracking, e-commerce events, or a cookie/consent banner tied to analytics. Covers installation, event tracking, user properties, e-commerce, Consent Mode v2 for EU/UK visitors, SPA pageview tracking, and BigQuery export. Skip for other analytics products (Plausible, Mixpanel, PostHog, Segment), Universal Analytics (sunset 2024), or server-side-only tracking.'

- "Why is this function failing?"

Quick reference for all GitNexus MCP tools, resources, and the knowledge graph schema.

'Core authentication patterns for Convex backends. Use when implementing auth checks in functions, storing users in database, creating auth helpers, or debugging auth issues. Works with any auth provider (Clerk, WorkOS, Auth0, etc.).'

description: Debug and audit financial models for errors — circular references, broken formulas, hardcoded overrides, balance sheet imbalances, cash flow mismatches, and logic gaps. Use when a model i

'Use this skill when building, explaining, or debugging regular expressions for pattern matching, validation, or text extraction. Trigger phrases: ''write a regex'', ''match this pattern'', ''validate this format'', ''extract from text''. Not for natural language parsing or full grammar parsing (use a parser instead).'

Use this skill when building, debugging, or optimizing CRM Analytics dataflows — defining node types (sfdcDigest, Append, Augment, computeExpression, computeRelative, Flatten, dim2mea, sfdcRegister), scheduling runs, handling run failures, and tuning performance. NOT for standard data processing outside CRM Analytics, for recipe-based data prep, or for SAQL dashboard query tuning.

Use this skill when configuring Salesforce B2B or D2C Commerce checkout: payment methods, shipping/tax integration, guest checkout, order summary setup, and CartCheckoutSession state orchestration. Trigger keywords: checkout flow, payment adapter, shipping tax integration, guest checkout, order summary, CartCheckoutSession, Managed Checkout, Commerce checkout flow. NOT for CPQ quoting, Checkout.com account management, or Service Cloud Order Management post-fulfillment logic.

Use this skill when designing the checkout experience for a Salesforce B2B or D2C Commerce storefront — covering runtime model selection (LWR Managed Checkout vs Aura Flow Builder), cart and line-item requirements, payment option scoping, shipping rule definition, and guest vs. registered buyer experience design. Trigger keywords: checkout flow design, checkout UX requirements, LWR vs Aura checkout, guest checkout design, payment options design, shipping rules, checkout experience planning. NOT for implementation: does not cover Apex adapter code, CartCheckoutSession debugging, Extension Point registration, or Flow Builder configuration steps.

Use this skill when writing or debugging SOQL relationship queries in Apex — child-to-parent dot notation traversal, parent-to-child subqueries, and polymorphic TYPEOF lookups. Trigger keywords: relationship query, subquery, dot notation, getSObjects, TYPEOF, WhatId, WhoId. NOT for aggregate queries (use apex-aggregate-queries), NOT for SOSL text search, NOT for Bulk API data loads (subqueries unsupported there).

Use this skill when building or debugging headless B2C Commerce storefront order submission using the Salesforce Commerce API (SCAPI) ShopAPI Orders endpoint or the legacy OCAPI /orders resource. Covers SCAPI order creation, SLAS authentication for shopper order access, OCAPI order placement and amendment, order status retrieval from the storefront layer, and notification/webhook configuration for order events. NOT for standard REST API (no generic CRUD against Order SObject). NOT for OMS Connect API (no OrderSummary, FulfillmentOrder, submit-cancel, or ensure-funds-async — those are covered by admin/commerce-order-management). NOT for CPQ order workflows.

Use this skill when building or debugging a custom Salesforce Commerce payment gateway adapter using the CommercePayments Apex namespace — covering adapter implementation, RequestType handling, PCI tokenization patterns, and integration path selection. NOT for billing (Salesforce Billing / blng namespace), NOT for configuring the legacy sfdc_checkout.CartPaymentAuthorize interface (see admin/commerce-checkout-configuration), and NOT for storefront UI payment form design.

Diagnosing and resolving common Apex runtime exceptions: NullPointerException, QueryException, DmlException, ListException, LimitException, TypeException. Use when debugging Apex runtime failures or writing defensive code. NOT for error handling framework design (use error-handling-framework). NOT for governor limit prevention strategies (use governor-limits). NOT for structuring try/catch blocks (use exception-handling).

Use when adding, configuring, or troubleshooting custom action buttons in the Salesforce CPQ Quote Line Editor (QLE), product configurator, or amendment/renewal screens. Trigger keywords: CPQ custom action, QLE button, SBQQ__CustomAction__c, CPQ Flow button, CPQ URL action. NOT for standard Salesforce quick actions on non-CPQ objects, Lightning Experience action bars outside the QLE, or CPQ Price Rules and Product Rules.

Use when building or debugging a browser-based CTI softphone adapter using the Salesforce Open CTI JavaScript API — covers callcenter.xml definition, Lightning utility item registration, core API methods (enableClickToDial, onClickToDial, screenPop, setSoftphonePanelHeight, saveLog), call logging as Task, and the lightning-click-to-dial LWC component. NOT for Service Cloud Voice Amazon Connect setup, Omni-Channel routing configuration, or CTI adapter AppExchange package selection.

Use when designing or implementing a custom logging framework in Apex: log sObject schema, log level gating, retention policies, batch purge jobs, and forwarding logs to external monitoring systems (Splunk, Datadog, etc.). NOT for built-in debug logs or Developer Console (use debug-logs-and-developer-console), NOT for exception capture and error propagation (use error-handling-framework), NOT for Event Monitoring (use security skills).

Use when diagnosing Apex behavior with debug logs, choosing log levels, or replacing `System.debug` habits with structured production logging and async job monitoring. Triggers: 'debug log', 'System.debug', 'logging framework', 'AsyncApexJob monitoring', 'production logs'. NOT for exception taxonomy design alone or for non-Apex observability platforms by themselves.

Use when building, debugging, or architecting a custom activity, custom split activity, or custom entry source for Marketing Cloud Journey Builder. Trigger keywords: custom activity, Postmonger, config.json, Journey Builder execute endpoint, customActivity.js, custom split outcome, installed package app extension, Journey Builder REST API. NOT for standard Journey Builder canvas configuration, out-of-the-box activity settings, Email Studio send activities, or general Marketing Cloud automation.

Complete reference for the Salesforce record save order of execution: all 18 steps from DB load through commit, covering trigger placement, validation rule sequencing, Flow execution timing, workflow field update re-fire behavior, and recursion patterns. Use when debugging unexpected automation behavior, designing multi-layer automation, or analyzing trigger vs Flow execution order. NOT for trigger framework design (use trigger-framework). NOT for record-triggered Flow patterns (use record-triggered-flow-patterns).

Use this skill when writing or debugging SOQL queries: SELECT syntax, WHERE filters, ORDER BY, LIMIT, OFFSET, relationship queries (child-to-parent and parent-to-child), aggregate functions (COUNT, SUM, AVG, MIN, MAX), and date literals. Trigger keywords: soql, query, SELECT FROM WHERE. NOT for SOQL security enforcement (use soql-security), query optimization and index tuning (use soql-query-optimization), or SOSL full-text search.

Designing and debugging Visualforce pages: standard/custom controllers, view state management, CSRF and SOQL injection security, PDF rendering, Visualforce email templates. Use when building custom UI pages or PDF outputs on the Salesforce platform. NOT for LWC development (use lwc/* skills). NOT for Visualforce email template syntax (use email-services).

>

> Track AI trading agents across Solana (pump.fun) and Monad (nad.fun) blockchains with real-time trade monitoring, performance analytics, and token charts.