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 mnemon-dev/mnemon --skill almagit clone --depth 1 https://github.com/mnemon-dev/mnemonWrote 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/mnemon-dev/mnemon/alma)<a href="https://agentmods.dev/skills/mnemon-dev/mnemon/alma"><img src="https://agentmods.dev/badge/skills/mnemon-dev/mnemon/alma/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/mnemon-dev/mnemon/alma"><img src="https://agentmods.dev/badge/skills/mnemon-dev/mnemon/alma.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 | $0.00068 | $0.00979 |
| Opus 5.5 | $0.00027 | $0.00392 |
| Sonnet 5.5 | $0.00014 | $0.00196 |
| Haiku 4.5 | $0.00007 | $0.00098 |
Grade A, and why
mnemon 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 6d 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 — 94 lines — stays where its author put it; the contents beside it link to each section on GitHub.
mnemon
mnemon is the source of truth. Alma's own memory/YYYY-MM-DD.md notes are a
per-session log, not the long-term store; durable facts belong in mnemon.
Use the shim, not the raw CLI
~/.config/alma/mnemon-bridge/mem is the only supported entry point. It pins the
store, keeps reads read-only, resolves short ids, and owns the MEMORY.md block.
Call it with Bash:
MEM=~/.config/alma/mnemon-bridge/mem
$MEM find "<query>" [--limit 5] [--chars 240] [--cat preference|decision|fact|insight|context]
$MEM save "<content>" --cat decision --imp 4 [--tags a,b] [--entities X,Y] [--source user|agent]
$MEM show <id-prefix>
$MEM related <id-prefix> [--edge causal|entity|semantic|supersedes|temporal]
$MEM recent [N]
$MEM status
Calling mnemon directly is only correct when the shim is missing. The raw CLI
needs a full UUID for show/related/forget, does not share the shim's store
pinning, and its recall --cat flag is accepted but ignored in 0.2.13.
Recall
Do not query on every turn. A recall costs 0.3–1.2 s and most turns do not need one. Recall when:
- the user asks "记不记得 / 之前 / 上次 / 我们说过 / do you remember"
- the topic switches to a specific project, service, repo, or host
- the user names something you do not recognise but that clearly comes from their history
- you are about to make a decision that may already have been made
Skip it for greetings, small talk, self-contained code questions, and any turn where the user already supplied the context.
Use the result silently. Do not narrate "I checked my memory".
Remember
Store one claim per entry, self-contained enough to make sense months later (with dates, file names, numbers). A conversation usually yields 0–3 durable entries; store nothing rather than padding the store.
| kind | example | --cat |
--imp |
|---|---|---|---|
| user preference | wording, taste, working style | preference |
4 |
| decided course | "use A not B, because …" | decision |
3–5 |
| durable technical fact | root cause, config shape, measured number | insight / fact |
3–4 |
| project background | what a module currently is | context |
3 |
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 6d ago First seen · 94 lines · 68 tokens per session scan A 19d01f3a18db
mnemon is a skill published in the GitHub repository mnemon-dev/mnemon (693 stars, last pushed 6d ago), licensed Apache-2.0. It adds 68 tokens to every session and 979 once invoked, about $0.0003 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 2026-11-01.
mnemon compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| mnemonmnemon-dev/mnemon | Skill | 693 | 6d ago | Apache-2.0 | A | 68 | Claude Code, Codex |
| innate-memoryvima-tech/Innate | Skill | 2 | 24d ago | MIT | A | 156 | Claude Code |
| innate-memoryvima-tech/Innate | Skill | 2 | 24d ago | MIT | A | 156 | Claude Code |
| brain-pagemindmuxai/brain.md | Skill | 594 | yesterday | Apache-2.0 | A | 47 | Claude Code |
| brain-setupmindmuxai/brain.md | Skill | 594 | yesterday | Apache-2.0 | A | 63 | Claude Code |
| brain-bootstrapmindmuxai/brain.md | Skill | 594 | yesterday | Apache-2.0 | A | 79 | Claude Code, Codex |
| brain-ingestmindmuxai/brain.md | Skill | 594 | yesterday | Apache-2.0 | A | 48 | Claude Code, Codex |