@garrytan
Writes software, dreams.
148 skills3 repositories 168.8K
Fix iOS bugs by editing source code, rebuilding the app, deploying it again, and verifying the result on a real device. Preserve a state snapshot as test data to help detect recurrence.
Launch GStack Browser, a visible Chromium browser controlled by an AI agent. Watch actions in a sidebar activity feed and use sidebar chat to request page navigation, clicks, form entry, and reading page content.
Extract information from web pages through Aside using existing signed-in sessions. Collect lists, tables, prices, or links and receive one JSON document, a structured data format, without changing the page.
Review a UI/UX plan before implementation, score its design dimensions, and identify missing decisions. Through discussion, the skill adds concrete improvements directly to the plan so the review produces an updated plan.
Connect a remote AI agent to your browser by generating a setup key and connection instructions. The connected agent receives its own tab, with an option to narrow its default page access.
Update gstack by detecting how it is installed, refreshing registered installations, and handling project copies and version migrations. Receive a summary of user-facing changes, along with installation-specific recovery information if an upgrade fails.
Turn approved mockups, design plans, or written descriptions into HTML and CSS using Pretext. Build dynamic layouts where text reflows and heights are calculated, choosing Pretext patterns suited to the design.
Test SwiftUI apps on a real iPhone or iPad by repeatedly capturing screens, analyzing them, taking actions, and verifying results. An embedded HTTP server enables interaction, with optional remote access through Tailscale.
Configure deployment, the process of publishing an application, for /land-and-deploy. Identify the hosting platform, production URL, health check endpoints, and status commands, then save those settings in CLAUDE.md for future deployments.
Investigate software errors and unexpected behavior through four stages: investigation, analysis, hypothesis, and implementation. It requires identifying the root cause before making fixes and distinguishes evidence from code inspection, tests, and live runs when reporting findings.
Merge an existing pull request, wait for automated checks and deployment, then verify production health with canary checks, targeted checks for problems. Continue the release workflow after /ship has created the pull request.
Check shell commands before they run and warn about destructive actions such as file deletion, database removal, and Git history changes. Users can approve ordinary warnings, while certain catastrophic command patterns are denied.
Audit code and configuration for security flaws, documenting attack paths, impact, evidence, and coverage gaps. Static review works without running the application; qualified runtime profiles add isolated reproduction and proposed repairs while keeping verification claims separate.
Find and fix bugs in websites, APIs, command-line tools, and background processes. Choose from three testing depths, verify fixes, and receive evidence of the changes plus a summary of readiness for release.
Create missing documentation for a feature, module, or whole project. Use Diataxis, a framework that separates tutorials, practical guides, reference information, and explanations, to organize documents around what readers need to learn or accomplish.
Turn stored knowledge pages into self-contained HTML files for sharing research, briefings, or travel plans. The workflow removes internal metadata and links, applies password protection by default, and offers optional cloud upload and expiring access links.
Validates skill files and registration consistency, then checks project tests and system health. Compares failures with recent changes to distinguish regressions, outdated tests, intermittent failures, and infrastructure issues, producing a report that tracks results over time.
Improve an existing skill’s instructions using benchmark tasks that measure its output. Generate starter tasks, preview costs, and keep version history, with validation checks deciding whether proposed edits can replace the current instructions.
Process large email, document, and transcript archives by routing items to AI models according to their importance. Rule-based privacy checks run before model calls, while extracted information becomes linked knowledge pages with progress checkpoints.
Adapt a knowledge collection's schema, the rules defining page types and relationships. Review untyped pages, propose new types, update classification settings, and apply those types to existing pages after previewing the changes and checking consistency.
Prepare a proven local skill for reuse in gbrain’s bundled collection. Guide removal of private details, generalize instructions and triggers, preview copied files, and verify the result before the user reviews and approves a commit.
Search shared organizational knowledge to answer questions about people, projects, decisions, and history. Save lasting conclusions as pages, update existing knowledge, and link related topics within the areas your account is allowed to access.
Turn voice calls into GBrain records containing transcripts, summaries, audio links, and references to people or companies. The workflow also sends summaries to a configured messaging service; automatic processing when calls end requires additional implementation.
Turn a feature into an agent skill or improve an existing one. Define purpose-specific quality criteria, review outputs with models from three providers, then add tests and check that user requests reach the intended skill.
Is this your account? Link GitHub to bring it onto your creator page (Japanese)