ClarifyPrompt MCP
A context-aware MCP prompt compiler that transforms vague prompts into platform-optimized prompts for 60+ AI platforms across 7 categories — grounded in your workspace signals (CLAUDE.md, AGENTS.md, .cursorrules, package.json), resolved intent, and the capabilities of the target model.
Send a raw prompt. ClarifyPrompt gathers the right context, resolves what you're actually trying to do, and returns a version specifically optimized for Midjourney, DALL-E, Sora, Runway, Higgsfield, ElevenLabs, Claude, ChatGPT, Cursor, or any of the 60+ supported platforms — with the right syntax, parameters, structure, and grounding.
New in 1.12.1: The real fix for #3 — thinking-channel models (gpt-oss, glm, …) now reliably produce optimized prompts instead of occasionally returning empty content. Root cause (re-investigated from scratch): they spend their token budget on the thinking channel first and never reach the final answer. The fix is a
max_tokensfloor for reasoning models (universal) plusreasoning_effort: "low"(for families that honor it, like gpt-oss; tune withLLM_REASONING_EFFORT) — not the previously-assumed/api/chatswitch, which turned out to be a dead end. Verified ongpt-oss:20b-cloudandglm-5.2:cloud(both 0% empty). See CHANGELOG.md.New in 1.12.0: Roadmap step #7 — ClarifyPrompt now speaks A2A (Agent-to-Agent). Set
CLARIFYPROMPT_TRANSPORT=a2aand it serves as a discoverable A2A peer: an agent card at/.well-known/agent-card.json, acompile-prompt-for-platformskill over JSON-RPC (message/send) with live SSE streaming (message/stream), and first-class task cancellation +input-requiredclarification round-trips — all powered by the same compose pipeline. Other agents can now call ClarifyPrompt to compile prompts. stdio stays the default; nothing about existing setups changes.






