activate_skill
Reactivate an ignored latent skill through the shared safe mutation pipeline.
LABSCO SUMMARY
Facts you can roll back sit beside a skill pipeline where a skill becomes a file only after something approved it.
Skills arrive through a funnel: evidence recorded, candidates read back, a decision taken, then generation — with the draft-and-publish route as the manual path to the same canonical state. Mutation is deliberately awkward where it should be. Purging facts and rolling back a scope refuse to run without an explicit confirmation, creating a skill touches no native target unless sync is asked for, and cleanup and repair stay inside targets the server itself manages rather than anything it finds. The recommendation tool accepts a dry-run flag for symmetry and is read-only either way, which is worth knowing before you build a plan around it.
What it is
A vendor-agnostic persistence layer for AI agents, held locally. It stores facts at project and global scope, turns repeated patterns into formal Agent Skills, and pushes approved rules and skills into the instruction files the tools you use already read — AGENTS.md, CLAUDE.md, .cursor/rules and their kin.
What you get
- A fact store with scope and visibility: facts persisted with tags, listed and filtered by scope and status, purged by ID or by whole scope behind a confirmation, and assembled into a Markdown context block sized to a character budget for injection into a system prompt (remember_fact, list_facts, purge_fact, context).
- Undo that is in place before you need it: snapshots listed per scope, the latest one restored behind a confirmation, and an audit trail of what changed (list_snapshots, rollback_scope, list_audit_events).
- Skills that start from evidence rather than a blank page — a recurring opportunity tracked with its evidence, candidates read back read-only, a decision recorded, then the physical skill generated once approved (track_latent_skill, recommend_skills, propose_skill, generate_skill, promote_skill_recommendation).
- A hand-authoring path beside it: an editable draft, validation against a draft, a canonical skill or a local path without touching files, and publication to canonical only once it passes (create_skill_draft, validate_skill, publish_skill, create_skill).
- Control over skills once they exist: listed and inspected for their metadata and triggers, edited, renamed while blocking unmanaged destination conflicts, activated or deactivated without deleting the SKILL.md, and shared into repository content (list_skills, get_skill_detail, update_canonical_skill, update_skill, rename_skill, activate_skill, deactivate_skill, share_skill).
- Adoption of skills already on disk, either imported into canonical storage or adopted in place without creating a duplicate slug (import_skill, adopt_skill).
- Sync that stays inside what it manages: canonical skills pushed to native targets with a drift decision, cleanup and repair limited to managed and orphan targets under a dry run, and approved rules pushed into the instruction files of registered hosts (sync_skills, cleanup_skill, repair_skills, sync_instructions).
- Project layout and health you can read before acting: the layout created, reported without mutation, or migrated from legacy content; and initialisation state, fact counts, active rules, registered skills, approximate size and host validations, alongside read-only environment diagnostics (initialize_project, inspect_project_layout, migrate_project_layout, status, doctor, host_setup, host_check).
Requirements
Python, and nothing else — no account, no key, and the store stays on your machine. It installs from PyPI and ships both a CLI and an MCP entry point from the same package, so the same operations are available either way. Pushing rules or skills into an editor or agent runtime additionally needs that host registered through host_setup under a host ID, with its own character and line budgets. Skills follow the agentskills.io structure, so a skill created here is a directory with a SKILL.md and its helper scripts rather than a proprietary record.
Setup effort
One command — uv tool install universal-memory
USE CASES
01
Carrying project decisions from one agent session into the next
Write it down once with its scope, then assemble it back into the prompt at a size the next session can afford.
02
Turning a workaround you keep re-explaining into a real Agent Skill
Record the recurrence with its evidence, review what came up often enough to formalise, then decide.
03
Publishing a skill you wrote by hand without shipping it unvalidated
Keep it a draft while it changes, validate without touching files, and promote only once it passes.
04
Keeping several editors aligned on the same rules
Approved rules are written into each registered host's own instruction file rather than pasted by hand.
05
Undoing a change that went wrong
Read the audit trail for the scope, then restore the latest snapshot with an explicit confirmation.
06
Bringing skills you already wrote into the store
Import or adopt the existing directory rather than recreating it and ending up with two slugs.
37 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
adopt_skill
Adopt an existing skill directory into UMEM without creating duplicate slugs.
cleanup_skill
Plan or apply managed-only native target cleanup for a canonical skill.
context
Assemble the active cognitive memory context for AI agent operations.
Retrieves and compiles project-specific or global context facts, preferences, and rules formatted in Markdown, optimized for injection into system prompts.
create_skill
Create a canonical Agent Skill without native sync unless sync is explicit.
create_skill_draft
Create an editable draft skill without canonical publish or native runtime writes.
deactivate_skill
Deactivate an active latent skill without deleting its physical SKILL.md.
doctor
Run read-only environment diagnostics for Universal Memory.
generate_skill
Generate the physical Agent Skill structure for an approved latent skill.
get_skill_detail
Inspect metadata and triggers for one registered skill.
host_check
Validate an agent host manifest without mutating files.
host_setup
Configure an agent host manifest through the safe mutation pipeline.
import_skill
Import an existing native or local Agent Skill directory into canonical UMEM storage.
initialize_project
Initialize the local Universal Memory project layout.
inspect_project_layout
Report the active project layout without mutating files.
list_audit_events
List audit events for a scope.
list_facts
List memory facts with optional scope and status filters.
list_skills
List registered skills and candidates without mutating local state.
list_snapshots
List created snapshots for a scope.
migrate_project_layout
Copy curated legacy project content into visible shared project storage.
promote_skill_recommendation
Promote an approved latent skill candidate into a canonical Agent Skill.
propose_skill
Review or decide a latent skill proposal.
If decision is omitted, returns the suggested name, scope, purpose, evidence, and explicit choices for a follow-up call.
publish_skill
Publish a validated draft as canonical and optionally sync native targets.
purge_fact
Purge one fact by id or all facts in a scope through the shared use case.
Requires confirm=True to execute.
recommend_skills
Review read-only latent skill recommendations captured by explicit skills.track evidence.
This never promotes, imports, syncs, creates, or mutates skills. dry_run is accepted for API clarity; non-read-only behavior is not supported.
remember_fact
Persist a memory fact through the shared safe mutation pipeline.
rename_skill
Rename a canonical skill slug while blocking unmanaged destination conflicts.
repair_skills
Plan or apply managed-only repair for orphan native skill targets.
rollback_scope
Rollback the latest created snapshot for a scope.
Requires confirm=True to execute.
share_skill
Explicitly copy a project skill into shared repository content.
status
Expose local cognitive persistence memory initialization and health check status.
Returns initialization state, fact counts, active rules, registered skills, approximate size, host validations, and recommended actions.
sync_instructions
Synchronize approved active rules into supported instruction targets.
sync_skills
Synchronize canonical Agent Skills into native targets with optional git warnings.
track_latent_skill
Explicitly track or increment recurrence for a latent skill opportunity.
update_canonical_skill
Update canonical skill content with validation and optional native sync.
update_skill
Update skill metadata or markdown through the shared safe mutation pipeline.
validate_skill
Validate a draft, canonical skill, or local skill path without mutating files.
BEFORE YOU INSTALL
What to know first
Tools you get
37 tools when we connected. The documentation advises calling bootstrap() once per session, but that tool is absent on v0.6.1.
Requires
Python 3.12+
Where it runs
The documented client runs the server on your machine.
CONTEXT COST
Moderate Standing cost
Rated by the size of the 37 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
TERMINAL
uv tool install universal-memoryMCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"universal-memory": {
"command": "uvx",
"args": [
"--from",
"universal-memory",
"umem-mcp"
]
}
}
}
