research-survey-visualizer
Survey research publications and create intuitive interactive visualizations inline in chat. Use when users ask to (1) survey/analyze a researcher's…
Skills
Find an Agent Skill for your task. Read the guide and copy the install command.
Updated 11-06 00:14
Writing & research
14,812 results
stacked-research
A research discipline that makes experiment results accumulate. Use it when designing, launching, and judging experiments, when referencing previous results, when handling the constitution, pre-registration, and verdict statements, when running the autoresearch loop, and when changing the canonical code and seeking approval. Re-buying the same conclusion, invalid levers (declaration exists but wiring does not...
talk-script-clarity
A skill for checking and polishing presentation scripts (speaker notes, lab seminar, conference, and defense scripts) so they can be understood in a single listen-through. 7 principles — (1) Plant a one-sentence theme at the beginning and the end, (2) carry one big metaphor from start to finish, (3) for each core concept, give the audience something to work through in their heads…
understand-knowledge
Analyze a Karpathy-pattern LLM wiki knowledge base and generate an interactive knowledge graph with entity extraction, implicit relationships, and topic…
writing-skills
Use when creating new skills, editing existing skills, or verifying skills work before deployment
domain-modeling
Build and sharpen a project's domain model. Use when discussing codebase terminology, writing or editing a CONTEXT.md, or recording or editing an ADR.
research
Investigate a question against high-trust primary sources and capture the findings as a Markdown file in the repo. Use when the user wants a topic researched,…
writing-for-agents
Writing documents for agents. Use when creating or editing skills, or modifying AGENTS.md or CLAUDE.md.
aikey-concept
AI KEY·Retell a concept — Retell a concept everyone is already using (pricing / topic selection / demand / conversion / customer acquisition…). Take it apart, and grow a ruler out of the cracks. Lock four things first, then write the draft, and finally finalize it through the person's own spoken delivery. Trigger: /aikey-concept, /retell-concept, "explain the concept of XX clearly", "what exactly is XX", "why does XX get more YY the more…
aikey-drama
AI KEY·Short-Drama Screenwriter — a vertical short-drama screenwriting skill. Given a topic or a piece of initial content, write a short-drama script that hooks from the opening and keeps viewers in at the end. Two kinds of output: ① a single-episode short drama (1–3 minutes, self-contained, something the account itself can shoot); ② a paid serial development package (a one-sentence selling point + synopsis + character bios + season-wide episode outline with paid unlock points marked + the first 3…
aikey-flow
AI KEY · Where will they swipe away — the "swipe-away points" special for spoken-word scripts. Listen to the script once on the audience's behalf: at each beat, what does the audience have in hand, what are they waiting for, and at which beat does it break — what they were waiting for never arrives, they already have it but you keep giving, or what you're giving can't be heard clearly. After diagnosing, by default proactively ask whether to do a marked-up revision. Not responsible for publishing red lines and whether it can be published (→ aikey-review), or whether the content hits home (→…
aikey-hook
AI KEY · The first five seconds — a specialized skill for short-video openings. First diagnose whether the content has substance (the gate), then compare the real-material entry point with the mass-suspense entry point, and produce candidates by combining opening moves and deliberate withholding, labeling each candidate with which principle it uses. For videos already shot and impossible to reshoot, pick the opening only from the exact words actually spoken. Not responsible for titles and cover headline text (→ aikey-title), for where in the full script viewers will swipe away (→ aikey-flow), or for whether it can be published (→…
aikey-mvp
AI KEY · Topic validation first — Topic MVP testing skill. Fragmented input (a one-line idea / chat log / X observation / candidate topic) → generate plain-text test content for posting on X (length unrestricted, write as long as the content needs, clear formatting) → register in the test pipeline →…
aikey-resonate
AI KEY · Does it hit anyone — resonance diagnosis skill. It does not start from theory but from the audience: what kind of takeaway does this piece of content leave the audience with (so that's how it is / I knew it / he said it for me), and will they act (save / share / comment / follow)? Every judgment points back to the original text and is labeled as evidence, inference, or to be verified. Mode A diagnoses your own draft; Mode B breaks down someone else's hit to reverse-engineer an angle you can borrow; Mode C…
aikey-review
AI KEY·Review once before publishing — a pre-publish review skill for spoken-word scripts. Review against the audience's four decisions: will they click in · will they stay · will they remember you · will they do something. First read the opening blind to judge stay-or-leave, then score information density sentence by sentence (60/80 score lines) + eleven questions + six reaction predictions (counting only the parts the audience can actually reach, including who the traffic is and how far from the business) + five red-line checks (fixes given in a safe version and a preserve-the-force version) +…
aikey-script
AI KEY·Spoken Script Writing — a collaborative spoken-script writing skill (primarily in the knowledge-paid spoken format). Select one of three script types + assemble from content units + co-create paragraph by paragraph; no one-click finished drafts. Trigger methods: /aikey-script, /写口播, /写稿, /aikey-写稿, "写个口播稿", "帮我把这个题写出来", "知识付费口播", "写条视频文案" Collaborative…
okf-validate
Validate OKF paths, frontmatter, required files, concept links, and citation structure deterministically. Use before consuming or distributing a bundle, while…
iblai-marketing-customer-research
When the user wants to conduct, analyze, or synthesize customer research. Use when the user mentions "customer research," "ICP research," "talk to customers,"…
garden-docs
Use when asked to verify docs match code, after structural changes needing doc updates, periodic doc maintenance, detecting stale docs, doc consistency checks,…
review-learning-repo-pr
Reviews pull requests that change the teaching structure of this repository. Use when PRs affect README.md, AGENTS.md, CLAUDE.md, .claude/, .github/, or…
validate-learning-repo
Audits this repository's teaching content against the real file structure. Use when checking README examples, AGENTS.md, CLAUDE.md, .claude/, .github/, or…
code-docs
Generate or update documentation from code analysis
humanizer-ru-local-edit
Use when a user asks to edit Russian prose that sounds formulaic or machine-generated while preserving its meaning and authorial voice.
humanizer-ru-local-edit-scope
Use when a user asks to edit Russian prose that sounds formulaic or machine-generated while preserving its meaning and authorial voice.