Menu
Skills in this repository
SkillsMP has collected 57 skills from koizumikento/stray-plugin. Open a skill to review its source and details.
koizumikento/stray-pluginShowing 17 of 57 collected skills.
Use when the user wants read-only triage of GitHub issues or pull requests and the next 1-3 maintainer actions. Do not use for comments, labels, assignments, closes, reviews, merges, CI repair, code changes, branch publishing, or any other GitHub write.
Use when global public patent databases are needed for prior-art, invalidity-candidate, FTO precheck, or landscape research across jurisdictions. Do not use for legal opinions, drafting, filing strategy, automation, or Japan-only J-PlatPat/FI/F-term research.
Use when the user wants a current, source-backed roundup of consequential Japanese news for a defined JST window in the fixed briefing format. Do not use for one-event deep dives, historical explainers, non-Japan roundups, or opinion-only coverage.
Use when a Japan-only patent search needs current J-PlatPat records, Japanese keywords, FI, or F-term for prior-art, invalidity-candidate, FTO precheck, or landscape work. Do not use for legal opinions, drafting, filing strategy, or primarily global searches.
Use when a provided page or URL should become clean, provenance-preserving notes or a permitted extraction with page chrome removed. Do not use for multi-source research, full-site crawling, or reproducing a third-party copyrighted page without permission.
Use when an ongoing task must be compacted so another turn or agent can continue with its files, decisions, evidence, failures, and next action intact. Do not use for one-off summaries, research briefs, or domain implementation.
Use when the user wants to design or compare a multi-agent execution model with roles, dependencies, handoffs, shared state, and quality gates. Do not use for single-agent work or authoring `.codex/agents/` files.
Use when the user wants to find and compare existing agent skills before deciding to adopt, adapt, or author one. Do not use for plugin-wide routing audits, implementation, generic web research, or package-manager workflows.
Use when the user wants to restyle an existing deck, document, report, HTML page, landing page, or demo while preserving content and structure, with format-native visual verification. Do not use for brand strategy, blank-slate creation, UI review, or…
Use when the user wants to define or refine a brand identity as a brief, territory, or mini style guide with actionable visual and verbal rules. Do not use for one-off artifact theming, product strategy, campaign copy, or final asset production.
Use when the user wants to create, edit, or repair a JSON Canvas 1.0 `.canvas` file with valid typed nodes, edges, IDs, and readable layout. Do not use for Mermaid, draw.io, generic whiteboards, ordinary JSON cleanup, or visual design work outside the JSON…
Use when the user wants to create or revise a conversion landing page in the current repository, including message, proof, CTA flow, search posture, and responsive implementation. Do not use for brand strategy, research-only work, product app flows, corporate…
Use when the user wants polished app screenshots for docs, landing pages, releases, demos, or social posts, including shot planning, capture-state sanitization, framing, export, and QA. Do not use for general UI review, redesign, browser-automation debugging,…
Use when the user wants operational knowledge or incident learnings turned into a maintainable SOP, runbook, checklist, or handoff guide for business or technical operators. Do not use for product specs, architecture decisions, implementation plans, or…
Use when the user needs a proposal, business case, plan, or memo asking a decision-maker for approval, funding, permission, or commitment, with evidence status and an approval path. Do not use for articles, operational runbooks, research-only work, or…
Use when the user wants a short animated GIF for a Slack emoji, message, or lightweight demo and needs loop design, crop, export, live constraint checks, and size optimization. Do not use for generic image editing, full video or motion-graphics production,…
Use when the user wants a test strategy, risk model, matrix, cases, QA/release checklist, regression scope, test data plan, coverage criteria, or automation recommendations as a design artifact. Do not use for test implementation, findings-first review, AI…