Performance optimization toolkit for OpenCode. Configs, memory skills, cleanup scripts, diagnostics, and comprehensive documentation — everything you need to make OpenCode faster in large projects.
File
Purpose
prompt.md
AI action file — paste this to any AI and it applies all optimizations automatically
templates/opencode.jsonc
Verified optimized global config
templates/opencode-project.jsonc
Verified optimized project config
File
Purpose
skills/persistent-memory/skill.json
Skill metadata
skills/persistent-memory/instructions.md
Agent memory read/write instructions
.opencode/memory/*.md
5 template memory files (decisions, patterns, solutions, progress, preferences)
File
Purpose
scripts/cache-cleanup.ps1
Windows cache cleanup (dry-run safe)
scripts/cache-cleanup-linux.sh
Linux/macOS cache cleanup (dry-run safe)
scripts/diagnostic.ps1
Windows performance diagnostics
scripts/diagnostic.sh
Linux/macOS performance diagnostics
File
Purpose
docs/performance-checklist.md
Quick-reference diagnostic checklist
docs/plugin-scores.md
Plugin evaluation table from 8 research reports
docs/known-bugs.md
15 documented issues with workarounds
docs/env-vars.md
Verified environment variables
docs/troubleshooting.md
Common errors and fixes
docs/install-plugins.md
Plugin install commands with version pinning
docs/graphify-setup.md
Graphify knowledge graph setup guide
docs/when-to-compact.md
Compaction decision guide
docs/mcp-optimization.md
MCP server audit and optimization
docs/memory-guide.md
Persistent memory usage guide
docs/wsl-notes.md
WSL-specific gotchas
docs/docker-setup.md
Docker setup guide
docs/ci-integration.md
CI/CD integration guide
Total: 14 documentation files
File
Purpose
examples/monorepo/opencode.json
Config for large monorepos
examples/small-project/opencode.json
Minimal config for small projects
File
Purpose
CHANGELOG.md
Version history
SECURITY.md
Security policy
CONTRIBUTING.md
How to contribute
tests/smoke-test.sh
Verify scripts work
Option A: Let an AI do it
Give prompt.md to any AI coding assistant. It will read your existing config, optimize it, and apply changes.
# 1. Copy optimized config
cp templates/opencode.jsonc ~ /.config/opencode/opencode.json
# 2. Copy project config
cp templates/opencode-project.jsonc ./opencode.json
# 3. Run diagnostics
bash scripts/diagnostic.sh
# 4. Clean cache
bash scripts/cache-cleanup-linux.sh --dry-run
All config fields verified against official OpenCode documentation:
Field
Value
Why
snapshot
false
Disables git indexing bloat on large repos
compaction.auto
true
Auto-manages context window
compaction.prune
true
Removes old tool outputs before compaction
compaction.reserved
20000-40000
Token buffer — higher = earlier compaction
watcher.ignore
["node_modules/**", ...]
Reduces file-watcher noise
provider.*.setCacheKey
true
Enables prompt cache affinity
autoupdate
"notify"
Prevents auto-download overhead
plugin
["name@x.y.z"]
Flat array, pinned versions
MIT