How szarkans/dont-forget works
don't forget is a Claude Code memory plugin backed by a user-owned vault of Markdown notes. The main workflow is saving a note, indexing it, and later searching text and wiki-link relationships; optional semantic search adds a meaning-based lane. Session and audit skills support review and thread tracking, while setup and health manage vault configuration and index status. The README documents several skill capabilities whose implementation details were not sampled; those relationships are shown only where documented.
Main entry points: index.py Vault indexer · search.py Text and link search · common.py Config and paths · SKILL.md Session review
Components 24
Plugin skills
- Save note
skills/that/SKILL.md - Recall notes
skills/about/SKILL.md - Session review
skills/session/SKILL.md - Vault setup
skills/setup/SKILL.md - Vault health
skills/health/SKILL.md - Vault audit
skills/audit/SKILL.md
Vault operations
- Markdown vault
- Note writer
scripts/vault-write.py - Vault discovery
scripts/setup.py
Index and search
- Vault indexer
scripts/index.py - Text and link search
scripts/search.py - Semantic runner
scripts/semantic.py - Embedding worker
scripts/semantic_worker.py - SQLite search index
- Vector index
Session and oversight
- Transcript reader
scripts/transcript.py - Thread tracking
scripts/threads.py - Vault audit
scripts/audit.py - Vault checkup
scripts/checkup.py - Context scan
scripts/hot-scan.py - Compaction guard
scripts/context-guard.py
Shared runtime
- Config and paths
scripts/common.py
Other components
- Developer
- Claude Code
Connections 29
- Developer → Claude Code: uses plugin
README.md - Claude Code → Save note: invokes
README.md - Claude Code → Recall notes: invokes
README.md - Claude Code → Session review: invokes
README.md - Claude Code → Vault setup: invokes
README.md - Claude Code → Vault health: invokes
README.md - Claude Code → Vault audit: invokes
README.md - Save note → Note writer: records note
- Note writer → Markdown vault: writes Markdown
scripts/vault-write.py - Vault setup → Vault discovery: configures vault
- Vault discovery → Config and paths: uses config
scripts/setup.py - Vault discovery → Vault indexer: builds index
scripts/setup.py - Vault indexer → Markdown vault: reads notes
README.md - Vault indexer → SQLite search index: builds SQLite index
scripts/index.py - Recall notes → Text and link search: searches vault
- Text and link search → SQLite search index: queries index
README.md - Text and link search → Semantic runner: requests meaning search
scripts/search.py - Semantic runner → Embedding worker: dispatches worker
scripts/semantic.py - Embedding worker → SQLite search index: reads chunks
scripts/semantic_worker.py - Embedding worker → Vector index: stores vectors
scripts/semantic_worker.py - Session review → Transcript reader: reads transcript
- Session review → Thread tracking: tracks threads
- Vault audit → Vault audit: reviews notes
- Vault health → Vault checkup: checks vault
- Compaction guard → Transcript reader: reads session data
- Context scan → SQLite search index: scans indexed notes
- Note writer → Config and paths: uses paths
scripts/vault-write.py - Text and link search → Config and paths: uses paths
scripts/search.py - Vault indexer → Config and paths: uses config
scripts/index.py