Jim Myhrberg
Community@jimeh · London, United Kingdom
I'm a software engineering mercenary, who plays with code, video games, and on occasion, design.
Agent Skills by Jim Myhrberg
Showing 64 vetted skills indexed across 1 GitHub repositories.
jimehhucode-upgrade-vscode
Replay Hucode patches onto a new upstream VS Code release branch.
jimehlaunch
Launch Code OSS builds with isolated profiles and unique debug ports for Playwright automation and dap-cli debugging.
jimehupdate-codex-sdk
Updates VS Code's bundled @openai/codex dependency and verifies the Codex Agent Host end to end.
jimehhucode-bump-version
Bumps Hucode release version and updates changelog from change fragments.
jimehgithub-copilot-upgrader
Upgrades the @github/copilot npm package and resolves resulting compilation and test failures.
jimehinit
Generate or update agent instruction files like AGENTS.md for AI coding agents.
jimehagent-customization
Create and troubleshoot VS Code agent customization files including instructions, prompts, agents, hooks, and skills.
jimehtroubleshoot
Diagnose unexpected chat agent behavior by analyzing Copilot debug logs in JSONL format.
jimehproject-setup-info-local
Scaffolds complete project structures for VS Code extensions, Next.js, Vite, MCP servers, and Python.
jimehcreate-prompt
Create reusable .prompt.md prompt files for repeatable AI tasks.
jimehproject-setup-info-context7
Scaffolds complete VS Code workspace projects using context7 library documentation.
jimehcreate-skill
Creates reusable SKILL.md files that package workflows into discoverable skills.
jimehinstall-vscode-extension
Installs VS Code extensions by ID using the workbench.extensions.installExtension command.
jimehcreate-instructions
Creates .instructions.md files defining project rules and coding conventions for AI agents.
jimehcreate-agent
Create custom agent definition files (.agent.md) for specialized AI assistant roles.
jimehchronicle
Analyze Copilot session history for standups, usage tips, cost analysis, and session search.
jimehget-search-view-results
Retrieves current search results from the VS Code Search view via command.
jimehcreate-hook
Create JSON hooks that enforce policy or automate agent lifecycle events.
jimehcode-oss-logs
Locate and read timestamped process logs from Code OSS dev builds.
jimehchat-customizations-editor
Guides development and testing of the Chat Customizations management editor in Code OSS.
jimehpolicy-and-managed-settings
Routes enterprise Copilot and VS Code controls to the correct policy implementation path.
jimehmemory-leak-audit
Audit TypeScript code for memory leaks and disposable lifecycle issues.
jimehcomponent-fixtures
Creates themed component fixtures for screenshot testing of VS Code UI widgets.
jimehauthor-contributions
Identify files an author contributed to on a branch, tracing contributions through git renames.
Frequently Asked Questions About Jim Myhrberg
FAQPage SchemaWhat tasks can I accomplish with jimeh's VS Code development skills?▼
You can upgrade Hucode against upstream VS Code releases, bump bundled Codex SDK versions, run unit/integration/smoke tests, fix CI failures, symbolicate native crash dumps, analyze heap snapshots and CPU profiles, create agent customization files, and validate UI scenarios with recorded video evidence.
Who are these skills designed for?▼
They target engineers working on the VS Code codebase itself: core workbench developers, Copilot Chat and agent host engineers, and release maintainers handling Hucode version bumps, upstream syncs, and patch replays. Some skills, like crash dump symbolication, are restricted to VS Code team members.
How do I run and debug VS Code from sources using these skills?▼
The launch skill starts Code OSS in a throwaway profile with unique debug ports, letting you drive the workbench with @playwright/cli while attaching a Node debugger via dap-cli. The vscode-dev-workbench skill covers running the vscode.dev server locally against microsoft/vscode sources.
What prerequisites and dependencies do these skills require?▼
Requirements vary: fix-ci-failures needs the gh CLI; flaky-smoke-tests needs Azure CLI; symbolicate-crash-dump requires macOS or Linux plus electron-minidump; UI driving skills need @playwright/cli; and most skills assume a local checkout of the microsoft/vscode repository with its dev build environment.
How do these skills support performance and memory investigations?▼
The chat-perf, auto-perf-optimize, heap-snapshot-analysis, cpu-profile-analysis, and memory-leak-audit skills cover benchmark runs, renderer heap snapshot capture and comparison, V8/Chrome profile analysis, retainer-path tracing, and disposable-pattern audits for event listeners and lifecycle callbacks.