recall-context
Show what memory files are loaded and available in the current session.
Search across all memory tiers (registers, daily logs, archive) for relevant information.
$ npx -y skills add davegoldblatt/total-recall --skill recall-search --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/recall-searchContext preview
The summary Claude sees to decide when to auto-load this skill.
Search across all memory tiers (registers, daily logs, archive) for relevant information.
description: Search across all memory tiers (registers, daily logs, archive) for relevant information. argument-hint: <query>
Search across all memory tiers for relevant information.
Query: $ARGUMENTS
Search in this order (narrow to wide):
Already loaded — check if the answer is here first.
Search all register files for content matching the query:
Search the last 14 days of daily logs. Start with most recent.
Only search here if nothing found in tiers 1-3, or if the query explicitly asks about historical/past items.
Use grep/ripgrep across the memory directory for matching content. This is keyword/pattern matching — search for relevant terms, synonyms, and related phrases.
Present results grouped by source, most relevant first:
Found [N] relevant entries: [registers/decisions.md] (confidence: high) 2026-01-15: Chose JWT over session cookies for stateless scaling [registers/tech-stack.md] (confidence: high) Auth: JWT tokens with 24h expiry, refresh via /auth/refresh [daily/2026-02-01.md] [14:32] Discussed adding OAuth2 for third-party integrations, decided to defer
Tell the user:
No matching entries found in memory. Searched: CLAUDE.local.md, [N] registers, [N] daily logs, [N] archive files.
Suggest they use `/recall-write` to capture the information if they have it.
A memory system for Claude Code that remembers what matters and forgets what doesn't.
Show what memory files are loaded and available in the current session.
Mark memory entries as superseded. Preserves history, does not delete.
Add durable IDs to memory entries. Required before /recall-maintain can run.
Scaffold the Total Recall memory directory structure in this project.
Pressure-based memory cleanup. Demotes, archives, or pins entries when working memory exceeds budget.