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 omermaksutii/mnemo --skill skillgit clone --depth 1 https://github.com/omermaksutii/mnemoWrote 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/omermaksutii/mnemo/skill)<a href="https://agentmods.dev/skills/omermaksutii/mnemo/skill"><img src="https://agentmods.dev/badge/skills/omermaksutii/mnemo/skill/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/omermaksutii/mnemo/skill"><img src="https://agentmods.dev/badge/skills/omermaksutii/mnemo/skill.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00050 | $0.00572 |
| Opus 5.5 | $0.00020 | $0.00229 |
| Sonnet 5.5 | $0.00010 | $0.00114 |
| Haiku 4.5 | $0.00005 | $0.00057 |
Grade A, and why
mnemo 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 1mo 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 — 53 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Mnemo — persistent memory
You have access to a persistent memory layer via these MCP tools:
mnemo_recall(query, k?, scope?)— semantic search across stored memoriesmnemo_remember(content, scope?, tags?)— capture a new memorymnemo_forget(id)— delete a memorymnemo_list(scope?, limit?)— browse recent memoriesmnemo_stats()— show memory engine status
When to use each
mnemo_recall — call this BEFORE answering anything where prior context might help:
- "what's our convention for X?"
- "how did we decide to handle Y?"
- "why is Z structured this way?"
- whenever you're starting a task in an unfamiliar area of the codebase
Use natural-language queries. Don't try to construct keyword searches — the index is semantic.
mnemo_remember — call this when:
- The user explicitly asks ("remember this", "save this for later")
- You make or affirm a non-obvious decision (capture it as
scope: "project") - The user states a personal preference that applies across projects (capture as
scope: "global") - You finish a task and notice something worth not having to rediscover
Default scope is "project" (auto-tied to the current repo).
mnemo_forget — only when the user explicitly asks to forget, OR when you discover a memory contradicts current reality and should be replaced.
Examples
User: "we always use Vitest, never Jest"
→ mnemo_remember({ content: "we always use Vitest, never Jest", scope: "global" })
User: "what test framework do we use?"
→ mnemo_recall({ query: "test framework preference", k: 3 }) then synthesize the answer.
User: "forget that pnpm thing"
→ mnemo_list({ limit: 50 }) to find it, then mnemo_forget({ id: "..." }).
Discipline
- Don't capture things that are already in CLAUDE.md or other always-loaded files (Mnemo is for the long tail that those files can't carry)
- Don't capture transient task state — capture lasting facts
- Prefer global scope for cross-project preferences; default to project scope otherwise
- When recall returns nothing useful, say so plainly — don't guess or hallucinate from low-similarity hits
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.
- 1mo ago First seen · 53 lines · 50 tokens per session scan A 6d6f541f7b79
mnemo is a skill published in the GitHub repository omermaksutii/mnemo (44 stars, last pushed 4mo ago), licensed MIT. It adds 50 tokens to every session and 572 once invoked, about $0.0002 per session on Opus 5.5. 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
qdrant-vector-search
High-performance vector similarity search engine for RAG and semantic search. Use when building production RAG systems requiring fast nearest neighbor search, hybrid search with filtering, or scalable vector storage with Rust-powered performance.
read-handoffs
Resume a session by loading recent handoffs from Engram. Use at the start of a session or when the user wants to review what was done previously.
handoff
Save a session handoff to Engram for future sessions. Captures decisions, todos, blockers, mental model, and next steps as a structured memory. Use when the user wants to preserve session context before ending work.
velesdb-memory
Use durable, explainable, self-improving memory across a coding session via the velesdb-memory MCP server. Trigger whenever the velesdb-memory MCP tools (remember/recall/recallfused/relate/why/feedback/forget/rememberextracted/extractionstatus/entity/memorystatus) are available and the work would benefit from…
velesdb-learning-loop
Turn every velesdb design / implement / verify cycle into durable, connected memory so the SAME mistake is never repeated and later decisions build on what was already learned, instead of re-deriving it or silently colliding with it. Trigger BEFORE starting a design or implementation on velesdb / velesdb-core /…
session-continuity
Use when an agent should pick work up from Engram quickly and leave behind a clean structured handoff for the next session. Best for multi-step coding, investigations, releases, debugging, and any task where recent decisions and open loops matter.