By doodledood
Optimize LLM prompts iteratively: review against 10-layer frameworks for clarity and issues, compress for token efficiency while preserving semantics, craft or refine from first principles via targeted discovery, and verify improvements with read-only agents on prompt files.
Verifies prompt compression quality. Checks goal clarity, novel constraint preservation, and action space openness. Flags over-specification and training-redundant content. Returns VERIFIED or ISSUES_FOUND.
Reviews LLM prompts against first-principles. Evaluates using the prompt-engineering skill's principles and reports issues without modifying files.
Verifies prompt token efficiency. In single-file mode, identifies inefficiencies (redundancy, verbosity). In two-file mode, verifies compression is lossless by comparing original vs compressed.
Iteratively auto-optimize a prompt until no issues remain. Uses prompt-reviewer in a loop, asks user for ambiguities, applies fixes via prompt-engineering skill. Runs until converged.
Compresses prompts/skills into minimal goal-focused instructions. Trusts the model, drops what it already knows, maximizes action space. Use when asked to compress, condense, or minimize a prompt.
Iteratively optimizes prompts for token efficiency by maximizing information density - reduces verbosity, removes redundancy, tightens phrasing while preserving semantic content. Use when asked to compress, shorten, reduce tokens, tighten, maximize density, increase information density, or make a prompt more concise.
Craft or update LLM prompts from first principles. Use when creating new prompts, updating existing ones, or reviewing prompt structure. Ensures prompts define WHAT and WHY, not HOW.
Review and analyze LLM prompts against prompt-engineering principles. Provides detailed assessment without modifying files.
Uses power tools
Uses Bash, Write, or Edit tools
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
First-principles workflows for Claude Code. Built by developers who understand LLM limitations.
Experienced developers frustrated by hype-driven AI coding tools. If you're tired of chasing the latest "game-changing" prompt that produces code you spend hours debugging, these plugins offer a grounded alternative.
Our approach:
Add the marketplace:
/plugin marketplace add https://github.com/doodledood/claude-code-plugins
Browse and install:
/plugin list
/plugin install vibe-extras@claude-code-plugins-marketplace
Multi-provider LLM consultations via Python/LiteLLM. Get perspectives from different models (OpenAI, Anthropic, Google, local models) without leaving Claude Code.
/review - PR review with severity-tagged findings/analyze-code - Architectural and security analysis/investigate-bug - Root cause analysis via external model/ask - Single-model consultation/ask-council - Multi-model ensemble (3 models in parallel)Requires Python 3.9+ and uv. Supports any provider via LiteLLM.
Standalone utilities for codebase exploration, web research, git operations, and code maintenance.
Research & exploration:
/explore-codebase - Structural overview with prioritized file list (quick/medium/thorough/very-thorough)/research-web - Deep web research with parallel investigators and multi-wave synthesisGit utilities:
/rebase-on-main - Safe rebasing with conflict resolution guidance/rewrite-history - Restructure commits for clean, reviewer-friendly PRsCode maintenance:
/clean-slop - Remove AI-generated noise (redundant comments, verbose patterns)/update-claude-md - Create or maintain CLAUDE.md project instructionsCode review:
/build-review-persona - Mine GitHub review history to generate a calibrated review-as-me skill/address-pr-comments - Triage and fix PR review comments with false positive detection/pr-merge-prep - Autonomously bring a PR to mergeable state (conflicts, CI, comments, description sync, re-review)/babysit-pr - Loop pr-merge-prep until PR is merge-ready, then ask before mergingSee also: manifest-dev for manifest-driven development—define acceptance criteria, let the verify-fix loop handle the rest.
Foundational documents for solo developers and small teams. Define once, reference consistently.
Customer & brand:
define-customer-profile - Create CUSTOMER.md (foundation for all other decisions)define-brand-guidelines - Communication style and voicedefine-design-guidelines - UI/UX patterns for your audiencedefine-seo-strategy - Traditional SEO + AI citation optimizationOther:
/audit-ux - Check UI changes against your design guidelinesCraft, analyze, and optimize prompts from first principles.
/prompt-engineering - Craft or update prompts from first principles (WHAT/WHY not HOW)/review-prompt - Analyze against 10-Layer Architecture framework (read-only)/optimize-prompt-token-efficiency - Iteratively optimize for token efficiency/compress-prompt - Compress into single dense paragraph for context injectionFrontend design patterns for distinctive, non-generic UI experiences.
scrollytelling - Scroll-driven storytelling with pinned sections, progressive reveals, and parallax effectsPersonal decision-making advisor. Understands your situation first, then researches and recommends.
/decide - Situation discovery → targeted research → decision framework → ranked recommendationAll writing skills in one place. Research-backed anti-AI prose, author voice capture, and voice-matched generation.
human-writing - Base knowledge with vocabulary kill-list, four-layer editing, craft fundamentals (auto-invoked for prose)/write - Full workflow: gather context → write → review → auto-fix → loop until cleancraft-author-voice - Capture your writing style into AUTHOR_VOICE.md/write-as-me - Generate content in your voice (requires AUTHOR_VOICE.md)Value-first promotion on social platforms. Finds high-signal posts and crafts replies that genuinely help conversations while naturally referencing your product.
/promote-on-x - Browse X feed, find relevant posts, craft structurally varied replies with writing review, post after approvalRequires Claude in Chrome extension and the writing plugin.
npx claudepluginhub doodledood/claude-code-plugins --plugin prompt-engineeringManifest-driven workflows for structured task execution with verification gates.
Post-processing utilities for manifest workflows. ADR synthesis from session transcripts.
Frontend design patterns and implementation skills for distinctive, non-generic UI experiences
Personal decision-making and life workflow support
Flexible multi-provider LLM consultations using Python/LiteLLM - includes consultant agent, review/bug-investigation commands, and consultant skill for deep AI-powered code analysis across 100+ models
Flexible multi-provider LLM consultations using Python/LiteLLM - includes consultant agent, review/bug-investigation commands, and consultant skill for deep AI-powered code analysis across 100+ models
Meta-cognition: refine input through brainstorming, refine output through challenge and condensed communication mode.
DevsForge Enterprise Prompt Optimization Architect delivering strategic prompt engineering methodologies, AI interaction optimization, and communication excellence frameworks
A comprehensive collection of custom skills for Claude Code covering engineering, product management, design, content, trading, database, QA, and educational domains
+ask +deep +web <- modifiers | optimize your prompts
Generate optimized prompts, PRDs, and task plans for Replit Agent that maximize AI understanding and minimize iterations