mnemosyne-oss/mnemosyne is a local AI memory layer that stores agent memories in SQLite instead of relying on cloud services. It is intended for coding agents and other agent frameworks, and can be used through MCP or a Python SDK. The catalogue skills connect agents to Mnemosyne’s memory features and related workflows.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add mnemosyne-oss/mnemosyne --skill mnemosyne-memory-overridegit clone --depth 1 https://github.com/mnemosyne-oss/mnemosyneWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/mnemosyne-oss/mnemosyne/mnemosyne-memory-override)<a href="https://agentmods.dev/skills/mnemosyne-oss/mnemosyne/mnemosyne-memory-override"><img src="https://agentmods.dev/badge/skills/mnemosyne-oss/mnemosyne/mnemosyne-memory-override/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/mnemosyne-oss/mnemosyne/mnemosyne-memory-override"><img src="https://agentmods.dev/badge/skills/mnemosyne-oss/mnemosyne/mnemosyne-memory-override.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00046 | $0.00595 |
| Opus 5 | $0.00023 | $0.00298 |
| Sonnet 5 | $0.00009 | $0.00119 |
| Haiku 4.5 | $0.00005 | $0.00060 |
Grade A, and why
mnemosyne-memory-override scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 60 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Mnemosyne Memory Override
CRITICAL RULE
NEVER use the memory tool for durable facts.
The legacy memory system is deprecated. It has a tiny 2,200 character limit
and does not use vector search. Mnemosyne is the single source of truth for
all durable user data.
Mapping
| Old (FORBIDDEN) | New (REQUIRED) |
|---|---|
memory(action="add", target="user", ...) |
mnemosyne_remember(content=..., importance=0.9+, source="preference") |
memory(action="add", target="memory", ...) |
mnemosyne_remember(content=..., importance=0.7+, source="fact") |
memory(action="replace", ...) |
Use the relevant Mnemosyne update flow: recall the current entry, then store corrected content with mnemosyne_remember; use an explicit update tool/flow when the integration exposes one. |
memory(action="remove", ...) |
Use the relevant Mnemosyne invalidation/forget flow: mark the obsolete fact invalid, supersede it with corrected content, or call the integration's forget/remove tool when available. Do not rely on passive aging for known-wrong durable data. |
When to use legacy memory (rare)
- Ephemeral session state (current todo list, temp flags)
- When Mnemosyne is explicitly confirmed down
Muscle Memory / Reflex Problem
You will reflexively reach for memory out of old habit. This is a known bug
in your training. Before EVERY memory tool call, pause and ask:
"Is this durable? Would I want this next session?"
If yes → use Mnemosyne
If no (temp flag, todo state) → memory is acceptable
Migration Cleanup
When moving data from legacy memory to Mnemosyne:
- Save to Mnemosyne first with
mnemosyne_remember - Then REMOVE the old entry from
memorywithmemory(action="remove") - This prevents stale duplicates and confusion
Enforcement
If you catch yourself typing memory(action= for durable data:
- CANCEL that tool call
- Use Mnemosyne instead
- Set importance >= 0.7 for anything that matters later
- If you already polluted legacy memory, remove the entry immediately
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 60 lines · 46 tokens per session scan A 1ad5e3199aa9
mnemosyne-memory-override is a skill published in the GitHub repository mnemosyne-oss/mnemosyne (3,082 stars, last pushed yesterday), licensed MIT. It adds 46 tokens to every session and 595 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
honcho
Configure and use Honcho memory with Hermes -- cross-session user modeling, multi-profile peer isolation, observation config, dialectic reasoning, session summaries, and context budget enforcement. Use when setting up Honcho, troubleshooting memory, managing profiles with Honcho peers, or tuning observation, recall…
qmd
Search personal knowledge bases, notes, docs, and meeting transcripts locally using qmd — a hybrid retrieval engine with BM25, vector search, and LLM reranking. Supports CLI and MCP integration.
llm-wiki
Karpathy's LLM Wiki: build/query interlinked markdown KB.
autocontext-consumer
Use when an agent needs to USE knowledge Autocontext already produced - find which scenarios have knowledge, read the playbook and lessons for one, understand the on-disk file and folder layout, and move knowledge between checkouts. Host-agnostic; requires only the autoctx CLI and the filesystem.
wikiskill-maintainer
Consolidate traces into the persistent wiki (WikiSkill).
mnemosyne-maintenance
Use when: upgrading Mnemosyne, diagnosing slow/hung consolidation (mnemosynesleep), fixing missing embeddings, or troubleshooting import/version mismatches.