searching-apple-noteslisted

# Apple Notes Semantic Search & Capture This skill gives Claude semantic search over the user's Apple Notes archive and the ability to write new notes back. It uses Qwen3-Embedding-0.6B embeddings (1024-dim, multilingual, strong on Chinese-English mixed content) and ChromaDB. Searches understand meaning and synonyms — searching for "burnout" finds notes about "exhaustion" and "压力" even when neither word literally appears. ## Path resolution The skill's helper scripts live next to this SKILL.md, in the `scripts/` subfolder. They figure out the parent repo's location automatically — there is nothing for the user or for Claude to configure. In the commands below, replace `SKILL_DIR` with the absolute path of this skill folder, which is wherever the user copied it (typically `~/.claude/skills/searching-apple-notes/` or `<project>/.claude/skills/searching-apple-notes/`). ## Prerequisite check Before the first search of a session, verify the index exists by running: ```bash python3.12 SKILL_DIR/scripts/vector_search.py --stats ``` If this returns an error about ChromaDB not being found, the user has not completed the setup steps in the parent repository. Tell them to follow the install instructions in the repo README and then come back. Do not proceed. ## Searching ```bash # Default: full-corpus semantic search, 10 results python3.12 SKILL_DIR/scripts/vector_search.py "query text" # Specify result count python3.12 SKILL_DIR/scripts/vector_search.py "query text" -n 20 # F