Search across all working memory for knowledge, decisions, people, or topics. Supports "how does X work", "what did we learn about Y", "when did we decide Z", "any gotchas with X", and "what's my P0 list". Returns matching entries from any node.
From session-memorynpx claudepluginhub brightwayai/session-memory-plugin/searchSearches issues by text query across title, description, and ID. Supports filters (--status, --label, --assignee, --type, etc.), sorting, limits, and JSON output.
/searchSearches Redpanda Connect components (inputs, outputs, processors, caches, rate-limits, buffers, metrics, tracers) matching the provided description.
/searchSearch AI experiments by query, tool, tag, rating, or date filters with sorting options and display formatted terminal results including stats.
/searchSearches pro-workflow learnings database using BM25 full-text search, returning top 10 ranked matches with category, mistake details, and application counts. Supports category and project filters.
/searchSearches your Notion workspace for pages, databases, and content matching the natural-language query, returning a scannable list with titles, types, descriptions, and clickable links.
/searchSearch Claude's persistent memories for content, patterns, or keywords via <query> [limit]. Returns matches with relevance scores, human-readable timestamps, and highlighted keywords.
Cross-project, cross-type search across all working memory.
| Query type | Examples | What to search |
|---|---|---|
| Knowledge | "how does their approval process work", "what do we know about onboarding" | MODEL, INSIGHT, LESSON, RECIPE entries |
| Gotcha | "any gotchas with Acme's procurement", "watch out for" | GOTCHA entries |
| Lesson | "what went wrong with the rollout", "what worked for retention" | LESSON, CORRECTION entries |
| Recipe | "how do we handle enterprise proposals", "process for onboarding" | RECIPE entries |
| Decision | "when did we decide on pricing", "why did we choose that vendor" | DECISIONS in LOGs |
| Person | "what do we know about Kim" | PEOPLE + LOG mentions |
| Topic | "everything about our hiring process" | All entry types |
| Artifact | "where's that proposal", "find the deck" | ARTIFACTS in LOGs |
| Blocker | "what's blocked" | BLOCKERS in SUMMARYs |
| Action | "what's my P0 list" | NEXT ACTIONS in SUMMARYs + LOGs |
| Question | "what questions are open" | QUESTIONS + OPEN THREADS |
Before searching: Check if ~/Documents/Claude/memory/ is accessible. If not, use the request_cowork_directory tool to request access:
mcp__cowork__request_cowork_directory(path="~/Documents/Claude")
The user will see an approval prompt. Wait for the mount to succeed before proceeding. If the user declines, explain that memory cannot be searched without this folder and stop.
~/Documents/Claude/memory/DASHBOARD.md to get the list of all nodesSearch all entry types: SUMMARY, LOG, INSIGHT, LESSON, MODEL, GOTCHA, RECIPE, CORRECTION, PEOPLE, SIGNAL.
Note per match: node, entry type, date, relevant snippet.
Prioritize knowledge entries over logs:
## What we know about: "[query]" — [count] entries across [node count] projects
### Mental Models
[node-id] ([date]): [MODEL entry]
### Lessons
[node-id] ([date]): [LESSON entry]
### Gotchas
[node-id] ([date]): [GOTCHA entry]
### Recipes
[node-id] ([date]): [RECIPE entry]
### Corrections
[node-id] ([date]): [old] → [corrected]
### Also in session logs
[node-id] ([date]): [LOG snippet]
Unified actionable list:
## All [Blockers/P0 Actions] — [today's date]
**[node-id]**: [description]
Group by node, relevance-sorted:
## Search: "[query]" — [count] results
### [node-id] ([match count])
**[date]** [type]: [snippet]
/learn or /note to capture now