Menu
Search Recall memory using FTS5 full-text search across all tables
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Add a structured memory record to Recall — breadcrumb, decision, or learning
Run health checks on all Recall memory subsystems — database, MCP, hooks, embeddings
Flush current session to Recall database and capture a Library of Alexandria entry
Browse and search the Library of Alexandria — curated knowledge entries with extracted wisdom
Show recent Recall memory records across all tables or a specific table
Scout an unfamiliar codebase — memory-first repo map, key paths, tests, risks, and next steps, with a strict sensitive-data boundary
| name | recall-search |
| description | Search Recall memory using FTS5 full-text search across all tables |
Prefer memory_search; use memory_hybrid_search for natural-language semantic retrieval. The CLI remains an explicit fallback.
The canonical workflow below remains authoritative, but equivalent behavior is not assumed across hosts.
Search across all Recall memory tables (messages, LoA entries, decisions, learnings, breadcrumbs) using SQLite FTS5 full-text search. Supports AND, OR, NOT operators, prefix matching, exact phrases, hard table filtering, and soft table biasing.
recall search "<query>"
Arguments:
<query> (required): Search query — use the user's argument textOptions:
-p <project> — Filter by project name-t <table> — Hard-filter to one table: messages, loa, decisions, learnings, breadcrumbs--bias-type <table> — Softly boost one table without filtering other matches. Same values as -t.-l <n> — Max results (default: 20)--show-provenance — Show Record Provenance for every result (by default only unknown provenance is flagged)# Search all memory
recall search "kubernetes deployment"
# Search decisions only (hard filter)
recall search "database choice" -t decisions
# Prefer decisions first, but still show matching learnings/messages/etc.
recall search "database choice" --bias-type decisions
# Filter by project
recall search "auth middleware" -p my-api
# Exact phrase with limit
recall search '"rate limiting strategy"' -l 5
# FTS5 operators
recall search "webpack OR vite" -t learnings
# Rule of thumb: use -t when you want ONLY that table; use --bias-type when you want that table first.