agentic-coding-tools

Synthesize implementation prompts for a fresh agent - auto-decomposes if task is too large

Multi-vendor adversarial review — parallel agents, skeptical verification, iterative consensus

Synthesize a complete implementation spec for a pair-programming session with a new coding agent to collaboratively build the discussed task

Reflect on conversation + codebase to ensure prompts capture everything before execution

Code review for uncommitted changes, staged changes, specific files, or a PR

Generate a walkthrough prompt and have an agent actively review it step-by-step

Walk through actual code in a codebase, reading files and showing what's really there, pausing for questions at each milestone

Visual UI testing and verification via browser automation. Use when testing user flows, checking styling/formatting, or investigating UI issues in the browser.

Analyze Claude Code transcripts (.jsonl). Use for searching history, extracting patterns, comparing sessions.

Web UI cloning specialist. Use when you need to create pixel-perfect HTML/CSS/JS reproductions of websites by visual observation. Analyzes target UIs via browser automation, extracts styling, generates code, and iteratively refines until matching. Handles sidebars, cards, lists, code blocks, diffs, and common UI patterns.

Holistic design critic that engages with a super-agent-driven walkthrough. Expects a walkthrough prompt file path as input. Spawns the walkthrough as a super-agent, then critiques the design at each milestone — catching under-engineering, over-abstraction, missing edge cases, and cross-cutting issues that the implementing agent couldn't see while building piece by piece.

Build a phased prompt chain with orchestration file from a monolithic implementation prompt

Extract and save browser screenshots from session transcripts

Wrapper around codex CLI for non-interactive runs (prompt via args/file/stdin) with resume support. Outputs session_id for conversation continuation.

Multi-agent debate orchestrator. Use when you want multiple AI perspectives on a question, plan review, or architectural decision. Runs parallel agent panels with a Chair synthesizing responses.

Wrapper around cursor-agent CLI for non-interactive runs (prompt via args/file/stdin) with resume support.

Wrapper around gemini CLI for non-interactive runs (prompt via args/file/stdin) with resume support. Outputs session_id for conversation continuation.

ALWAYS use this skill instead of Read when reading Claude Code transcript files (.jsonl in ~/.claude/projects/, ~/.claude/history.jsonl, or paths containing 'transcript'). Token-efficient format that extracts USER/ASSISTANT exchanges and tool summaries, strips metadata.

Use ONLY when user explicitly mentions saving plan to ".ai-reference" or "ai-reference" or "for reference". Phrases like "save this plan to ai-reference", "save plan for reference", "archive this plan". Do NOT use for generic "save this plan" requests.

SDK-based Claude agent with full Task tool access. Use for executing handoff prompts or spawning nested sub-agents.

This skill should be used when the user asks to "implement this plan", "generate execution prompts", "super implement", "create handoff from plan", or wants to transform a large plan into self-contained execution-ready prompt artifacts.