Use this skill when the user asks to audit a website for SEO issues — meta tags, sitemap, robots.txt, broken links, headings, duplicate titles, keyword density. Also use when asked about SEO analysis, website health checks, or running the seo-audit CLI or MCP tools. This skill covers running audits via CLI (npx…
Design world-class SaaS landing pages grounded in real references from saaspo.com (via the saaspo MCP). Use when the user asks to design or build a landing page, marketing site, or any page for a product — especially when they provide copy and want it designed "like a senior designer", or mention saaspo, design…
Forensic retrospective on how work actually FLOWED through the kanban-pro board — reconstruct each card's timeline from the change-log (lane moves, bounces, forced moves, attention, dispatch loops, rounds, cycle time), attribute it to the model that really served it (model-pro requested-vs-resolved, context overflow…
Use when working with the OpenCTI threat-intelligence platform via the opencti MCP tools — explains OpenCTI's STIX data model, entity types, indicators vs observables, relationships, containers, TLP markings, confidence, connectors/ingestion, and RBAC, and maps each concept to the right MCP tool. Load before creating…
Consolidate the brain — refresh the index, review sessions since the last reflection, distill durable facts into memory files, compact the store, propose skills, commit. Use when asked to "reflect" or "consolidate memory"; also invoked headless by the nightly job.
One-shot sweep of Reddit and X (formerly Twitter) reactions to a product launch, announcement, release, or news moment in a time window, read out as volume, representative quotes, themes, and notable accounts. Use whenever the user asks how people are reacting, what the sentiment or reception is, whether a launch…
Analyze a chart — set up symbol/timeframe, add indicators, scroll to key dates, annotate, and screenshot. Use when the user wants technical analysis or chart review.
Produce fact-checked academic research drafts using the Aurelius MCP tools. Use whenever the user asks to research a topic, write a paper/literature review/long-form report, generate a LaTeX document, draw an architecture or flowchart diagram, or verify claims and citations against real scholarly sources. Screens the…
Trigger: web-research-mcp, checkreference, saveresearch, cache miss, resolvereference. Delegate WebSearch/WebFetch+summarize to a Haiku 4.5 subagent instead of the main thread's model.
A command for collecting technical documentation from a framework's website or a supplied URL and saving the original data. A framework is a reusable foundation for building software.
Use whenever the user asks Claude to inspect or control their Hyprland desktop via the hyprland-mcp server — moving/resizing/closing windows, switching or reorganizing workspaces, listing or reconfiguring monitors, taking screenshots, sending notifications, reading or tweaking live Hyprland config, or running…
Generate a technical design doc (problem statement, approach, architecture diagram, trade-offs, open questions) from the real technical work done in this session, and save it into the project as Markdown and HTML. Use when the user runs /design-doc, with optional flags --activate KEY or --help.
Verify football match events and result readiness with Proofline evidence, signed x402 proof packets, and the latest Injective testnet settlement commitment. Use when an agent must check a score, event, final result, paid proof, or settlement gate without confusing scheduled, delayed, replay, and live data or…
Verify AI-generated educational visuals for science, engineering, medical education, charts, diagrams, and technical illustrations using structured specs, extraction tools, domain rules, and evidence-backed findings.
Design Newton cloth animation scenes with particle resolution, thickness, anchors, distance constraints, collision objects, and external fields. Use when a prompt involves fabric, sheets, flags, curtains, nets, or other thin deformable surfaces.
Use CupOracle to answer FIFA World Cup 2026 questions (fixtures, live scores, consensus odds, knockout bracket) and to buy a vetted, CLV-audited edge that the agent pays for ITSELF via x402 on Injective — under a spend cap, always citing the on-chain receipt. Invoke when the user asks about a World Cup match, "who's…
★not rated 0 2d agoA98 tokens
originalMIT
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: