Yusong-Enceladus

Autonomously improve a generated paper via GPT-5.4 xhigh review → implement fixes → recompile, for 2 rounds. Use when user says "改论文", "improve paper", "论文润色循环", "auto improve", or wants to iteratively polish a generated paper.

Autonomous research review loop using any OpenAI-compatible LLM API. Configure via llm-chat MCP server or environment variables. Trigger with "auto review loop llm" or "llm review".

Autonomous multi-round research review loop using MiniMax API. Use when you want to use MiniMax instead of Codex MCP for external review. Trigger with "auto review loop minimax" or "minimax review".

Autonomous multi-round research review loop. Repeatedly reviews via Codex MCP, implements fixes, and re-reviews until positive assessment or max rounds reached. Use when user says "auto review loop", "review until it passes", or wants autonomous iterative…

Generate and rank research ideas given a broad direction. Use when user says "找idea", "brainstorm ideas", "generate research ideas", "what can we work on", or wants to explore a research area for publishable directions.

Workflow 1: Full idea discovery pipeline. Orchestrates research-lit → idea-creator → novelty-check → research-review to go from a broad research direction to validated, pilot-tested ideas. Use when user says "找idea全流程", "idea discovery pipeline", "从零开始找方向",…

Validate a research idea BEFORE implementation. Tests core assumptions with minimal experiments. Kills bad ideas in hours, not weeks. MANDATORY before any implementation work begins.

Generate publication-quality figures and tables from experiment results. Use when user says "画图", "作图", "generate figures", "paper figures", or needs plots for a paper.

Decide whether one new intake item (a dictation, meeting transcript, pasted text, document or screenshot text) is about the same concrete object as one of a few pre-retrieved candidate events, starts a new event, belongs to no matter at all (stays unfiled),…

Write the short title and the single "现在到哪一步" (where things stand, as of the event's latest item) line for one event, plus 1-4 status facts that each carry a progress state (planned / in_progress / done / cancelled / info) and cite the item ids they come…

Summarize one file the user dropped into the app (document, spreadsheet, slides, PDF or scan, e-mail, calendar invite, contact card, e-book, archive, saved web page, code or data file) after the organizer has already extracted its text: write one line saying…

Score how important each event is for the home screen right now (importance 0..1 plus a short Chinese reason), so the app can show one default order that mixes importance and recency. Use when a batch of items has just been organized and the organizer passes…

Read one image the user dropped into the app (chat screenshot, chart or dashboard, slide, handwriting or whiteboard photo, receipt or invoice, scanned page, label / shipping form / sign, or anything else with text) in two steps: name its type, then extract…

Cut one long item that may cover several separate matters (a meeting transcript, a long dictation or note, a brainstorm dump) into contiguous parts, one per matter, so each part can be filed into its own event. Use when the organizer passes an item's text as…

[P1 stub] Read-only outward recall for other agents: answer "what is the state of X in the user's life" by returning matching events (title, 现在到哪一步, cited facts, item ids) from the local organizer. Use only when another local agent, with the user's…

Use when a PowerPoint .pptx deck needs click-by-click entrance animations, grouped column/card/flow reveals, semantic block animation, animation QA, or animation XML/API troubleshooting. Adds entrance animations to existing slide elements while preserving the…