Menu
SkillsMP has collected 56 skills from liujiayi3703/scientific-research-skills. Open a skill to review its source and details.
Showing 40 of 56 collected skills.
Use when the user wants Bayesian intrinsic-growth valuation for a company, including implied growth, 3-5 year growth probabilities, FOMO versus fundamentals, TAM, market share, margin trajectory, valuation multiples, industry cycle, or probability updates…
Use when the user provides a ticker or listed company and wants a source-backed buy-side equity memo, investment committee note, deep stock analysis, target-price scenarios, SEC/IR financial analysis, industry chain, competition, valuation, catalysts, risks,…
Use when the user provides a ticker or asks for GF-DMA scoring, valuation/trend health, moving-average support, price-to-DMA divergence, ATR divergence, escape risk, estimate revisions, healthy momentum, or whether a rising/falling stock is fundamentally…
Use when the user explicitly asks for an OS-level desktop or system screenshot, such as full screen, a native app window, or a pixel region, and no tool-specific capture or app-testing capability is the better fit.
Use when the user wants Serenity-style alpha research that converts market-moving news, product launches, procurement signals, supply-chain changes, or earnings-call clues into investable hypotheses, small-cap beneficiaries, financial-statement transmission…
Use when the user provides a ticker or asks for TAM-Adj-PEG valuation, growth-stock cheap/expensive judgment, TAM-supported runway, quality-adjusted PEG, or classification as core growth, high-beta growth, turnaround, option-like, or cyclical exposure.
Use when interacting with or testing local web applications using Playwright, especially localhost apps that need clicking through flows, capturing page screenshots, inspecting browser console errors, debugging UI behavior, or verifying frontend behavior.
Use when the user asks for academic research tasks: deep research, literature review, systematic review, meta-analysis, research question refinement, academic paper drafting or revision, citation checks, peer review, editorial decisions, research-to-paper…
Use when the user requests code-generated art, generative art, algorithmic art, p5.js sketches, seeded randomness, flow fields, particle systems, or interactive visual parameter exploration.
Use when the user asks to create a poster, visual art, static design, PNG/PDF design artifact, print piece, cover, composition, or typography-led visual deliverable. Use for original visual design even when the final export is PDF.
Use when code imports `anthropic` or `@anthropic-ai/sdk`, or when the user asks for Claude API, Anthropic SDK, Managed Agents, prompt caching, tool use, batch, files, citations, memory, model migration, or cache-hit debugging. Skip provider-neutral, OpenAI,…
Use when the user wants to co-author documentation, proposals, technical specs, decision docs, structured content, writing plans, draft refinement, reader-focused review, or iterative document collaboration.
Use when the user wants to create, read, edit, format, convert, or inspect Word documents (.docx), including reports, memos, letters, templates, headings, page numbers, tables of contents, images, tracked changes, comments, find-and-replace, or polished Word…
Use when the user asks how claude-mem works, what it is doing, how observations are captured, when memory injection happens, where data lives, or how persistent memory is added to sessions.
Use when writing internal communications such as status reports, leadership updates, 3P updates, company newsletters, FAQs, incident reports, project updates, team announcements, or executive summaries.
Use when the user asks for 妈妈模式, 妈妈唠叨, mama mode, 唠叨模式, /pua:mama, /pua mama, or $pua-mama; applies a Chinese-mom nagging tone while preserving the same quality bar.
Source text: Chinese
Use when the user mentions Nano Banana, nanobanana, AI image editing, image generation, prompt-to-image iteration, image-to-image refinement, product mockups, visual concept generation, or local Nano Banana plugin assets.
Use when the user writes, revises, reviews, or plans a Chinese NSFC grant application, National Natural Science Foundation proposal, guo-zi-ran, 国自然, 基金本子, 立项依据, 科学问题, 研究内容, 技术路线, 创新点, 可行性分析, or grant review response.
Use when setting up, configuring, or troubleshooting OpenClaw agent memory, memory plugins, gateway memory, persistent observations, memory slots, worker startup, observation feeds, or OpenClaw install commands.
Use when asked to find the ideal path, map a codebase into feature-grouped flowcharts, identify duplicated concerns across features, unify duplicated systems, or audit architecture before a refactor.
Use when the user needs PDF document processing: read, extract text/tables/images, merge, split, rotate, watermark, fill forms, encrypt/decrypt, OCR scanned PDFs, inspect PDF layout, or create utility PDFs. Do not use when PDF is only the export format for a…
Use when a PowerPoint file is involved as input or output: .pptx, slide deck, pitch deck, presentation, speaker notes, templates, layouts, comments, extracting text, editing slides, creating decks, combining files, or splitting presentations.
Use when the user asks for PUA Pro platform extensions: self-evolution tracking, compaction-state protection, KPI reporting, leaderboard, ranking, weekly report, promotion report, 段位, 周报, 述职, 排行榜, 自进化, or /pua:pro features. Do not use for exact alias commands…
Source text: Chinese
Use when the user invokes $pua-cancel-loop or /pua:cancel-loop to cancel the current PUA autonomous loop cleanly.
Use when a task has failed repeatedly, the agent is stuck tweaking the same approach, suggests manual work too early, blames the environment without verification, or the user says try harder, stop giving up, figure it out, or similar. Also useful for complex…
Use when the user invokes $pua-flavor or /pua:flavor to list or switch PUA flavor options.
Use when the user asks for pua-ja mode or Japanese 詰め style after repeated task failures, same-approach tweaking loops, passive debugging, unverified environment blame, or frustration such as もっと頑張れ / なんとかしろ.
Source text: Japanese
Use when the user invokes $pua-kpi or /pua:kpi to generate a PUA KPI report.
Use when the user asks for /pua:pua-loop, $pua-loop, PUA loop mode, 自动循环, 一直跑, or autonomous iterative development that continues until independent verification says the task is done.
Source text: Chinese
Use when the user invokes $pua-mama or /pua:mama as the alias form of mama mode.
Use when the user invokes $pua-off or /pua:off to turn off PUA mode for the session.
Use when the user invokes $pua-offline or /pua:offline for offline PUA mode, no-network debugging, or local-only task execution.
Use when the user invokes $pua-on or /pua:on to enable PUA mode, turn on PUA enforcement, or reactivate strict task execution.
Use when the user invokes $pua-p10 or /pua:p10 for P10 CTO mode with PUA enforcement.
Use when the user invokes $pua-p7 or /pua:p7 for P7 senior engineer mode with PUA enforcement.
Use when the user invokes $pua-p9 or /pua:p9 for P9 tech lead mode with PUA enforcement.
Use when the user invokes $pua-pro or /pua:pro for the PUA Pro alias command.
Use when the user invokes $pua-reap-orphans or /pua:reap-orphans to clean up orphaned PUA agents or processes.
Use when the user invokes $pua-survey or /pua:survey to inspect the current PUA environment and status.
Use when the user invokes $pua-team-status or /pua:team-status to report active PUA team agents and their state.