Load prior Zettelkasten cards into agent context via memex index reads and searches.
npx skills add https://github.com/iamtouchskyer/memex --skill memex-recall
Please help me install this Agent Skill. Skill: memex-recall Source: https://github.com/iamtouchskyer/memex/tree/main/skills/memex-recall Command: npx skills add https://github.com/iamtouchskyer/memex --skill memex-recall
This Skill prevents agents from starting work with no context by loading relevant prior knowledge from a Zettelkasten-style memex repository at the start of every new conversation or task, reducing redundant effort and preserving continuity across sessions.
Before starting the task, run memex read index and then perform targeted memex searches and reads to summarize relevant cards for context.
High-intent search queries and answers about installing and using this skill.
To load prior Zettelkasten memory into an agent context, run memex read index first, then perform targeted memex searches and reads to summarize relevant cards for continuity before task initiation.
The best way to retrieve prior notes and prevent zero-context starts is applying index-first recall, which prefers a curated keyword index as the fastest entry point to relevant cards for context-rich responses.
Controlled link traversal works by following linked cards selectively with guardrails of max_hops=3 and max_cards_read=20, avoiding runaway reads while gathering sufficient memory context for the task.
Yes, memex recall requires access to the memex CLI or MCP memex_read and memex_search tools to enforce guardrails and retrieve relevant cards via index reads and keyword searches.
Yes, you can use bilingual queries for searching Zettelkasten memory, generating 2–3 search queries using both English and Chinese when appropriate to read matching cards directly if the index is insufficient.
iamtouchskyerSave distilled development insights into persistent Zettelkasten memory cards.
23blocks-OSRetrieve prior discussions and decisions via memory-search.sh before starting new tasks.
cajiasDesigns Amazon S3 Vectors semantic search and RAG workflows for AWS-native architectures.
roy2392Retrieve long-term memory via vector search and recursive graph traversal on SQLite.
