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 agentmods add skills/cbuntingde/kimi-memory/kimi-memorynpx skills add cbuntingde/kimi-memory --skill kimi-memorygit clone --depth 1 https://github.com/cbuntingde/kimi-memoryWrote 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/cbuntingde/kimi-memory/kimi-memory)<a href="https://agentmods.dev/skills/cbuntingde/kimi-memory/kimi-memory"><img src="https://agentmods.dev/badge/skills/cbuntingde/kimi-memory/kimi-memory.svg" alt="Measured on agentmods" 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 | $0.00054 | $0.02569 |
| Opus 5 | $0.00027 | $0.01285 |
| Sonnet 5 | $0.00011 | $0.00514 |
| Haiku 4.5 | $0.00005 | $0.00257 |
Grade A, and why
kimi-memory 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 today.
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 — 92 lines — stays where its author put it; the contents beside it link to each section on GitHub.
kimi-memory
The kimi-memory plugin gives the Kimi agent a three-layer memory store backed by local SQLite, exposed through MCP tools, and seeded at session start.
The skill follows progressive disclosure: this file covers routing, hygiene, types, and the typical flow. Deeper material lives in references/:
references/tools.md— the full MCP tool catalog.references/recall-acknowledgement.md— how to acknowledge[recall],[focus],[thread],[tool-recall]segments on the hook status line.references/active-memory.md— v9+ behaviour: continuous retrieval, mid-turn recall, decay, cross-session thread, background consolidation, auto-GC.references/decay-contract.md— the Ebbinghaus decay + reinforcement contract, with the formula and the migration that introduced it.
Layer routing rules
Pick the right scope for every fact you save or recall:
| Layer | Scope | When |
|---|---|---|
| Global user memory | scope: "global" |
Cross-project preferences, profile facts, reusable procedures (e.g. "user prefers dark mode", "use pnpm globally", "how to onboard the agent on a new machine"). |
| Project durable memory | scope: "project" (default) |
Repository-specific facts, decisions, conventions, workflows for the active project. |
| Project working memory | working_memory_set/get/clear (project-only) |
Transient current focus, in-flight tasks, recent decisions you want surfaced next turn. |
| Session archive | conversation_* tools (project-only) |
Automatic full per-session transcript; no manual reads unless the user asks. |
What ships with it
4 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.
- today Changed · +4 lines 8b3c4c39a332
- 3d ago First seen · 88 lines · 54 tokens per session scan A 83a87738ac1c
kimi-memory is a skill published in the GitHub repository cbuntingde/kimi-memory (0 stars, last pushed today), licensed MIT. It adds 54 tokens to every session and 2,569 once invoked, about $0.0003 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 2026-08-31.
Other skills, from other repositories
membrain
Shared long-term memory over MCP. Call memorycontext once at the start of any session/task to load what is already known; use searchmemory before answering questions about the user or their projects; save durable facts, preferences, and decisions with savememory / savememories. All agents and the user share the same…
delx-memory
Local-first agent memory MCP. No API keys. Prefer MCP tools if connected; otherwise the package CLI. Use when the user wants Delx Memory data or actions through an agent.
plur-memory
Persistent learning for AI agents. Open engram format. Your agent learns from corrections, remembers across sessions, and transfers knowledge across domains.
memorix-memory
Use when prior workspace context, past decisions, solved bugs, handoff state, or durable project knowledge would help a coding task.
plur-session-end
Extract durable learnings at the end of a session. Saves corrections, preferences, and codebase patterns as engrams — nothing ephemeral, nothing sensitive.
memorix
Use when Claude Code needs Memorix shared memory, reasoning, Git Memory, mini-skills, session handoff, orchestration coordination, or integration troubleshooting.