majiayu000/spellbook / skills/vscode-doctor
vscode-doctor
VerifiedDiagnose slow or freezing VS Code-compatible editors with evidence-first, zero-hardcoded-assumption workflow. Use when the user reports editor lag, typing delay, UI freezes, extension host stalls, file watcher noise, high editor CPU/RSS, uses VS Code/Cursor as a file browser over a large folder, or wants a safe editor performance audit.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 21.7 KB |
| evals/README.md | markdown | 2.0 KB |
| evals/evals.json | config | 7.3 KB |
| scripts/collect_vscode_diagnostics.shexecutable | script | 4.0 KB |
evals/README.md
View on GitHub# vscode-doctor Eval Plan
## Goal
Verify that the skill diagnoses editor performance without hardcoded local paths, fixed root-cause rankings, fixed extension blocklists, fixed generated-directory lists, or fixed benefit percentages.
## Core Checks
- Uses the collector with case-specific inputs instead of assuming defaults.
- Ranks findings only from evidence present in the current prompt/output.
- Treats operating-system, renderer, watcher, extension, and cache issues as candidate classes, not as guaranteed causes.
- Uses placeholders for paths and commands in examples.
- Avoids applying changes until the user confirms.
- States when impact cannot be quantified because baseline data is missing.
- Separates file-browser mode from normal coding mode when a broad parent directory is open.
- Explains `files.watcherExclude`, `search.exclude`, and `files.exclude` with their editing/search/visibility tradeoffs.
- Treats Python, Ruff, Git, and Copilot setting changes as optional experiments with explicit feature costs, not default fixes.
- Defaults to a plain-language report for non-technical users and moves jargon-heavy evidence into a technical details section.
- Offers a small set of reversible choices instead of a long technical options table.
- When commands or file edits were performed, reports a user-visible action log: what was checked, what changed, why it changed, what the user might notice, side effects, verification, and rollback.
- Treats collector output as sensitive and redacts local paths, project names, command lines, and extension identifiers before external sharing.
- Routes explicit large-workspace, file-browser, narrow-workspace, extension-profiling, and maintenance intents before proposing fixes.
- Keeps whole-machine CPU load as context instead of using it as the editor diagnosis.
## Manual Eval Prompts
The prompts live in `evals.json`. They intentionally use placeholders and generic categories so the skill must ask for or discover the missing context.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| rea-analysis-optimizationmorluto/reaOptimize REA providers and shared analysis workflows for agent context, latency, CPU, and memory while preserving correctness and Evidence. Use for analysis bottlenecks and cache or query design; CI/test-suite optimization belongs to rea-ci-tests-audit. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-ci-tests-auditmorluto/reaAudit REA CI and test suites for correctness, fixture drift, false passes, and measured performance; use for suite or workflow investigations and efficiency changes, rather than ordinary feature testing or a single test failure. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| rea-tool-designmorluto/reaAudit, investigate, design, or change REA investigation tools, CLI/MCP contracts, provider capabilities, and Evidence semantics. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-auditDietrichGebert/ponytailQuality audit of the whole repo: bugs, security, real load, missing tests, speed, and what to delete. Most important first. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |