Pine Script Pro
Authors production-grade Pine Script v6 for TradingView (indicators, strategies, libraries). Use when the user asks to "write a Pine Script", "build a TradingView indicator", "create a strategy", "code an oscillator/overlay/screener", "Pine Script v6", "TradingView alert", or describes any trading idea they want on a chart ("alert when RSI crosses 70 on the 4h", "show me a divergence indicator",…
Ballot Advisor
Researches the voter's ballot and recommends how to vote in local and state elections by matching candidates and ballot measures to the VOTER'S stated values. Use when the user says "help me vote", "who should I vote for", "research my ballot", "voter guide", "candidate recommendations", "sample ballot", "explain these ballot measures", mentions an upcoming local/state/primary/general election, o…
Desktop Ui Craft
Architects and builds componentized, reusable, elegantly styled desktop application UIs — design-token theming, full interaction-state matrices (hover/pressed/focus/selected/disabled), rich well-featured components, and senior-designer restraint. Use when the user asks to build, design, restyle, extend, or polish any desktop app UI - "create a component", "build a panel/dialog/table/settings scre…
Davinci Resolve
World-class DaVinci Resolve support and post-production mentorship. Use whenever the user mentions "DaVinci Resolve", "Resolve", "BlackMagic editor", or asks about post issues that live in Resolve — choppy/laggy playback, "media offline", a project that won't open or crashes, color grading and node trees, color management (RCM, ACES, DaVinci Wide Gamut), Fusion compositing, Fairlight audio, codec…
Arch Audit
Analyzes an entire codebase for architectural smells, spaghetti patterns, tangled dependencies, and structural debt — prescribes concrete refactors grounded in Ousterhout, Fowler, Hickey, Feathers, and connascence theory. Use when the user asks to "audit the architecture", "review the codebase", "find architectural issues", "analyze tech debt", "find spaghetti code", "architectural smells", "why…
Plan Eval
Evaluates software implementation plans for soundness, completeness, best practices, and executability using parallel expert agents. Use when user says 'review this plan', 'evaluate the plan', 'check this plan', 'is this plan good', 'plan review', 'audit this plan', 'sanity check the plan', 'critique this plan', or wants feedback on any plan/design document. Also triggers when user shares a plan…
Arch Conformance
Discovers a codebase's intended architecture (from docs, ADRs, naming, package structure, and dependency direction), then deploys parallel agent teams to find code that violates it — layer skips, reverse dependencies, cycles, bypassed boundaries, leaked cross-cutting concerns, mutant duplicates. Writes findings to plan/*.md files. Use when the user asks to "find architecture violations", "check a…
App Harden
Audits an application for runtime stability, resilience, and attack-surface issues — runaway log growth, unbounded resource use, missing timeouts, exposed ports, SSRF, CSRF, leaking secrets, crash loops, and more. Use when the user asks to "harden", "stability review", "make production-ready", "audit for runtime issues", "security review", "find resource leaks", "check for DoS vectors", "pre-flig…
Refactor Pro
>
Pattern Architect
Reorganizes messy, tangled, or hard-to-extend code by applying the pattern that fits the smell — the Gang of Four catalogue plus stateful/messaging patterns (actors, managers, event buses, observers, state machines, queues). Use when the user wants to organize or untangle code, fix spaghetti, make something extensible, stop editing the same switch/if-ladder for every new case, split a class that…
Rust Test Separate
Separates inline #[cfg(test)] modules from Rust implementation files into dedicated test files. Use when user says 'separate tests', 'extract tests', 'move tests to separate files', 'split tests from implementation', 'reorganize rust tests', or asks about test file organization in Rust. Do NOT use for writing new tests, running tests, or non-Rust codebases.
Rust Uplift
Analyzes Rust codebases for higher-level abstraction opportunities: iterator chains, actor patterns, type-state, derive macros, ecosystem crates. Finds where idiomatic patterns reduce verbosity and improve maintainability. Use when user says 'improve Rust code', 'simplify', 'find abstractions', 'make idiomatic', 'reduce boilerplate', 'uplift', 'review Rust quality', or asks about modernizing Rust…
Fsharp Uplift
Analyzes F# codebases for higher-level, idiomatic abstraction opportunities: pipelines and composition, partial application, Option/Result combinators, computation expressions (result/asyncResult/validation, comprehensions, query), collection HOFs (choose/fold/groupBy), discriminated unions that make illegal states unrepresentable, active patterns, units of measure, exhaustiveness hardening, and…
Sprint
End-to-end workflow that takes a feature request (or any coding task) and drives it from framing to a machine-verified, adversarially-reviewed diff — pausing only at the two decisions that are the human's (approve the plan, accept the result). Use when the user says 'sprint', '/sprint', hands over a feature request, or wants the full plan-through-verified-change arc. The upgraded successor to dev…
Rust Warn Fix
Analyzes all Rust compiler and Clippy warnings, then intelligently fixes or silences each one using expert judgment. Use when user says 'fix warnings', 'clean up warnings', 'resolve clippy', 'zero warnings', 'handle warnings', 'warning cleanup', or when Rust build output shows warnings. Also use proactively when cargo build/clippy output contains warnings during development. Do NOT use for adding…
Bug Hunt
>
Plan Execute
Implements code changes from a plan markdown file with verification. Use when user says 'execute this plan', 'implement the plan', 'follow this plan', 'run plan-execute', or provides a .md plan file to implement. Reads plan files (standalone or index linking sub-files), verifies each item exists before fixing, logs false positives to ignored.md. Use exclusively for executing existing plans, not c…
Skill Forge
Researches any domain in depth, then builds a production-quality Claude Code skill encoding expert-level knowledge. Use when user says "forge a skill", "skill-forge", or invokes /skill-forge. Do NOT use for general "create a skill" requests unless user explicitly mentions skill-forge.
Rust Ci Green
Makes Rust projects CI-green by fixing all cargo fmt and clippy issues and ensuring GitHub Actions workflows produce green badges. Use when user says 'make CI green', 'green badge', 'fix CI', 'CI passing', 'fix the build', 'fmt and clippy', 'cargo fmt clippy', 'make it pass CI', 'workflow green', or when a Rust project's GitHub Actions are failing due to formatting or lint issues. Also use when s…
Canon R6 Mark Ii
Canon EOS R6 Mark II hybrid mirrorless expert — stills and video. Use when the user mentions "R6 II", "R6 Mark II", "EOS R6 Mark II", or asks how to configure, operate, or troubleshoot the camera (Canon Log 3, HDR PQ, 4K60 vs 4K30, oversampling, 40fps electronic burst, RAW Burst pre-shooting, focus bracketing/depth composite, anti-flicker vs high-frequency anti-flicker, IBIS, Movie Digital IS, su…
Plan Feature
Plans implementation of a software feature using parallel research agents, draft-then-critique workflow, and codebase-grounded vertical slices. Use when user says 'plan a feature', 'plan this feature', 'implementation plan', 'feature plan', 'plan the implementation', 'how should I implement', 'design this feature', or wants a technical plan for new functionality. Do NOT use for executing existing…
Code Scrutiny
>
Rust Commit
Ensures a Rust codebase is CI-clean before committing: auto-fixes formatting, clippy warnings, verifies tests and doc warnings so GitHub Actions badges stay green. Use when user says 'rust commit', 'make it green', 'CI ready', 'prepare to commit', 'clean commit', 'pristine commit', or after finishing Rust code changes and wanting a verified commit. Also triggers when user wants to verify Rust cod…
Canon C50
Canon EOS C50 cinema camera expert and general cinematography mentor. Use when the user mentions "C50", "EOS C50", "Cinema EOS C50", asks how to configure or operate the camera (C-Log 2/3, Cinema RAW Light, XF-AVC/XF-HEVC, dual base ISO, sub recording, proxies, anamorphic desqueeze, false color, waveform, assignable buttons, Custom Picture, Frame.io, ACES, ND filters with the C50), or asks genera…
Skill Refine
Audits and improves existing Claude skills against Anthropic's official best practices — third-person descriptions, concise SKILL.md bodies, progressive disclosure, appropriate degrees of freedom, no voodoo constants, one-level references. Use when the user says "improve this skill", "refine this skill", "audit this skill", "upgrade this skill", "review my skill", "check this skill against best p…