lpclaude-config

Analyze project's commands, subagents, and workers to identify needed conversions and optimizations

Audits and optimizes CLAUDE.md files against current best practices, extracting inappropriate content to external docs

Add parallel execution capabilities to existing Claude Code commands

Create new Claude Code commands with best practices and automatic validation

Optimize slash command definitions for best practices and performance

Translate Claude Code commands to Gemini CLI format

Validate existing Claude Code commands against best practices

Create comprehensive session documentation for seamless handoff to future agents

Create comprehensive project documentation capturing current session context and implementation strategy

Audits and optimizes repository README.md to ensure an outstanding first-impression user experience

Intelligently verify completion claims using parallel completion-verifier agents with smart task splitting

Intelligently group changed files into logical commits with semantic messages and push to origin

Analyze and rewrite Git commit messages based on the scope: "$ARGUMENTS"

Run 7 parallel security audits and consolidate findings by priority

Generate a comprehensive, trackable implementation plan from the current conversation context. Plans are saved in `docs/implans/` and serve as living documents for project execution.

Resume work on an existing implementation plan with flexible execution control

Verify and update implementation plan status based on actual project state

Natural language project improvement system that understands contextual improvement requests and creates targeted scanners with safeguards against infinite loops

Process job queue files (.md and .parallel.md files) from a specified folder with context-aware execution

Review current work context to identify remaining tasks that need to be captured as jobs, then create them using the addjob agent

Add specific insights to CLAUDE.md or extract learnings from conversation, placing them in the appropriate location

Audit agentic frameworks for outdated SlashCommand tool workarounds and identify migration opportunities

Parse the provided arguments to extract:

Convert PDF to Markdown using marker tool with minimal modifications - preserves original content

Analyze session context to identify and document issues encountered for systematic fixing

Answer questions without modifying any files

Safely simplifies a single tracked file in a git repo to its most essential form without losing functional information. One-pass, dry-run guarded, idempotent.

Create new Claude Code skills with best practices and proper directory structure

Optimize subagent definitions for reliable automatic invocation and peak performance

Audit subagents for compatibility, overlap, and optimization opportunities

Execute task instructions autonomously until completely done with full verification

Answer VS Code settings questions using local documentation

Apply a subtle color tint to VS Code workspace for visual differentiation

Execute a worker template from the workers directory

Creates job files for deferred task execution using the addjob utility. Invoke when users say "add a job", "create job", "create jobs", "make a job", "make jobs", "defer this task", "do this later", "schedule this", "queue this work", "save this for later", or any variation involving the word "job" or "jobs". Use when tasks need to be scheduled for later processing, when work should be deferred rather than executed immediately, when creating batch processing queues, or when user asks to create any job files. Automatically triggered by keywords like "job", "jobs", "defer", "later", "schedule", "queue", "batch", "todo", "task list", or "save for later". This agent uses the addjob bash command to create structured job files in the project's jobs/ directory. Use proactively when job-related keywords are detected, including plural forms.

Meticulous completion verification specialist that rigorously verifies whether claims about task completion are actually true. Expert at identifying incomplete implementations, missing tests, stubbed code, and premature completion declarations. Use ONLY when explicit verification is requested or definitive completion claims are made about deliverables. Invoke when user explicitly asks to "verify completion", "check if implementation is done", "confirm feature is complete", or makes declarative statements like "X is finished/implemented/ready". Do NOT trigger on planning discussions, status questions, or casual mentions of completion words. Use proactively when actual verification or validation is clearly needed.

Expert LLM delegation specialist that seamlessly connects to external language models including GPT-4, GPT-3.5, Gemini, Ollama, and local models via the llm bash tool. Invoke this agent to delegate tasks, get second opinions, validate solutions, or continue existing model conversations. Use when users mention external LLMs, API models, delegation, validation with other models, checking work, comparing responses across models, or need specific model capabilities. MUST BE USED PROACTIVELY when detecting phrases like "use gpt", "ask gemini", "check with", "delegate to", "second opinion", "validate with", "compare using", "continue conversation", "external model", "api model", or references to specific providers.

An expert documentation specialist that comprehensively audits and updates project documentation for accuracy, relevance, and consistency with the current codebase. Invoke this agent when documentation is outdated, inconsistent, or needs comprehensive review after code changes or refactoring. Use when you need to identify documentation gaps, fix inaccurate information, consolidate redundant docs, or ensure documentation reflects the current state of the project. This agent systematically analyzes code-to-documentation alignment, validates examples and instructions, and produces a detailed audit report with all necessary corrections. Use proactively when appropriate.

Expert code security and quality reviewer specializing in detecting technical debt, shortcuts, hardcoded values, personally identifiable information embedded in code or documentation that is not in .gitignore, brittle implementations that could cause bugs or security vulnerabilities. Use proactively when reviewing new code, investigating inconsistent bugs, auditing authentication systems, or preparing for code reviews to catch dangerous shortcuts early. Analyzes production code for hardcoded secrets, magic numbers, brittle conditionals, disabled safeguards, and workflow bypasses that indicate hacky implementations. Invoke when code needs security audit, when features work inconsistently, or when investigating technical debt before refactoring. Use proactively when appropriate.

Expert implementation plan auditor for reviewing implans for completeness, correctness, and compliance with requirements. Use proactively when users mention implan auditing, plan validation, checking for incomplete implementations, stubs, TODOs, missing test coverage, or when they need to verify implementation completeness. Analyzes both implan structure and associated codebase to detect gaps, placeholder code, untested functionality, and deviations from requirements. Outputs comprehensive audit reports with corrective actions and creates specialized testing phases for critical untested areas.