Use memloom memory well during any task. Recall before starting, save decisions and preferences as they happen, resolve conflicts. Use in any session where the memloom MCP tools are available.
Scanned 9/11/2026
Install to Claude Code
npx -y skills add memloom/memloom --skill memloom-use --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Memloom Use?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/memloom-memloom-use)More formats (shields.io, HTML) on the badges page.
---
name: memloom-use
description: Use memloom memory well during any task. Recall before starting, save decisions and preferences as they happen, resolve conflicts. Use in any session where the memloom MCP tools are available.
---
# Using memloom
memloom is the user's persistent memory across all their agents. You have its MCP tools:
`save_memory`, `recall_memory`, `read_passage`, `memory_history`, `list_conflicts`,
`resolve_conflict`, and schema tools. Habits that make it useful:
## Recall first
Before starting a task, recall what is already known about it: the project, the people,
the constraints. One `recall_memory` query with the task's key nouns is usually enough.
Results marked `context` come from the user's ingested files and carry a source citation;
treat them as quotes from the file, not as beliefs.
If a hit is truncated, `read_passage` with its id returns the full text.
## Save as you go
Save at the moment something becomes true, not at the end of the session:
- A decision was made: save it as a `fact` or `episode` with the why.
- The user corrected you or stated how they like things done: save a `preference`.
- You worked out reusable steps (a release ritual, a fix recipe): save a `procedure`.
Write memories as short, standalone sentences that will make sense months from now with
no session context. Include names, dates, and numbers; skip session-local details.
Do not save what the repo already records (code, git history, README facts) or secrets of
any kind.
## Respect the belief pipeline
A save can come back as `added`, but also as a new version of an existing memory or as a
conflict with one. That is normal. If the result names a conflict, tell the user; they can
resolve it in the viewer, or you can `list_conflicts` and `resolve_conflict` when they ask.
Every resolution is reversible, so resolving is never destructive.
## When recall seems wrong
Say what you found and where it came from rather than silently trusting or discarding it.
Memories reflect what was true when saved; the user decides what is current. `memory_history`
shows a belief's version chain when it matters.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!