Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/oneworks-ai/appnpx agentmods add skills/oneworks-ai/app/oneworks-memWrote 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/oneworks-ai/app/oneworks-mem)<a href="https://agentmods.dev/skills/oneworks-ai/app/oneworks-mem"><img src="https://agentmods.dev/badge/skills/oneworks-ai/app/oneworks-mem/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/oneworks-ai/app/oneworks-mem"><img src="https://agentmods.dev/badge/skills/oneworks-ai/app/oneworks-mem.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 | $0.00045 | $0.02063 |
| Opus 5.5 | $0.00018 | $0.00825 |
| Sonnet 5.5 | $0.00009 | $0.00413 |
| Haiku 4.5 | $0.00005 | $0.00206 |
Grade A, and why
oneworks-mem 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 — 107 lines — stays where its author put it; the contents beside it link to each section on GitHub.
当任务发生在 channel 会话里,或用户要求保存、读取、整理长期上下文时使用这个 skill。oneworks mem 是给 agent 在 shell 中调用的 CLI,不是发送到聊天频道里的文本命令。
在 channel 会话中,默认认为 oneworks mem 已经由环境注入可用。不要为了确认是否存在而先执行 which oneworks、oneworks --help 或 oneworks mem --help;直接按下面示例调用,只有命令失败、参数不确定且示例不足,或用户明确要求时才查询 help。
在 channel 会话中,不要让模型自己抄写平台 id。oneworks mem 会从环境变量和 server 写入的当前消息上下文文件读取 session id、channel id、群聊当前发送者 sender id 等元数据。群聊里多人轮流说话时,-s user 默认指向当前这条消息的发送者。
当前实体的 MEMORY_POLICY.md 定义“什么值得记、什么不应记、如何表述和何时淘汰”;entity.yaml 的 memory 段定义可写 scope、证据、有效期、敏感数据和每轮预算等硬边界。语义策略和硬边界都高于本 skill 的通用建议。命令拒绝写入时,不要改用别的文件路径绕过策略。
快速命令
oneworks mem get:读取当前默认记忆文件,等价于当前 channel id 下的README.md。oneworks mem patch "内容":向当前默认记忆文件追加内容。oneworks mem set "内容":覆盖当前默认记忆文件,只有明确要重写时使用。oneworks mem list:列出当前 scope 下已有记忆文件;不传-p时列出全部路径。
所有 subcommand 都支持:
-p, --path <path>:指定或过滤 id 下的相对文件路径。get/set/patch默认README.md,list不传时不过滤路径。-c, --channel <channel>:指定或过滤 channel,例如wechat。-f, --filter <id>:指定或过滤平台相关 id。对get/set/patch是目标 id,对list是过滤条件。-s, --scope <scope>:记忆维度,支持entity、channel、conversation、user、session、global。
Scope 选择
entity:用于当前实体跨 ChannelLink 复用的长期经验、操作约定和稳定判断。channel:默认 scope。用于当前群聊、私聊、帖子或平台会话的长期上下文;适合频道主题、项目背景、群约定、常用配置、长期排障线索。conversation:用于稳定对话线程的背景、决策和后续动作;新的物理 ChildSession 会继续加载。user:用于当前发送者个人相关信息;适合姓名、称呼、职责、稳定偏好、常用工作方式。只在信息来自本人、被明确确认,或对任务持续有用时写入。群聊中不要手填 sender id,让 CLI 从当前消息上下文解析。session:用于当前物理 ChildSession 的临时工作记忆;不会作为跨轮连续记忆自动加载。global:用于跨频道也成立的通用事实或用户明确要求全局记住的规则;谨慎使用。
什么时候读取
优先在这些场景读取相关记忆:
- 用户提到“之前”“上次”“按老配置”“记得我说过”“继续那个问题”等跨轮上下文。
- 任务需要知道频道长期背景、平台 id、项目约定、常见故障、用户偏好或之前的决策。
- 当前请求含糊,但已有记忆可能决定正确做法。
- 在群聊或私聊中遇到不熟的人、昵称、群内梗、项目名、表情含义、图片语境或关系距离,且这些信息会影响回复语气或任务判断时,先读小本本,不要只靠猜。
- 用户问“你知道 X 吗”“你还记得 X 吗”,或你准备吐槽/接梗但不确定上下文时,先
oneworks mem get/oneworks mem get -s user/oneworks mem list。 - 准备写入前,先
get或list,避免重复、冲突或把同一主题写到多个地方。
运行时加载结构化记忆时会先匹配可见性 group。orgs 必须命中;conversationTypes、entities、rooms、channels 等维度在同一 group 内按 OR、不同 group 之间按 AND 组合。不要因为实体名相同,就假设另一个组织、群聊或会话中的记忆可见。候选选择和预算也会按 visibility group 分配,避免单一高频群聊挤占全部上下文。
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 · 107 lines · 45 tokens per session scan A 6a973a34f74d
oneworks-mem is a skill published in the GitHub repository oneworks-ai/app (20 stars, last pushed 1mo ago), licensed MIT. It adds 45 tokens to every session and 2,063 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
neo4j-agent-memory-skill
Authoritative reference for the neo4j-agent-memory Python package — a graph-native memory system for AI agents built on Neo4j — and for the hosted service (NAMS) at memory.neo4jlabs.com. Use this skill whenever the user mentions neo4j-agent-memory, agent memory with Neo4j, context graphs, the POLE+O model…
wiki-ingest
Use when the user wants to ingest external sources into the vault — PDFs, markdown files, transcripts, Readwise exports, or any document. Triggers on "ingest", "process document", "import into vault", "distill this", "add to vault", or "extract knowledge from".
cross-linker
Use when the user wants to find and insert missing wikilinks across vault notes. Triggers on "cross-link", "find missing links", "weave links", "unlinked mentions", "add wikilinks", or "link notes together".
lov-auto-context
Manual or hook-triggered context operator. Evaluates the current session for pollution (long conversations, topic drift, stale noise), AND takes concrete context-shaping actions: writing project memory, updating global or project CLAUDE.md (with diff + confirm), and recommending harness commands like /fork, /compact…
lov-search-chat
A local search tool for recalling relevant parts of earlier AI conversations from an indexed memory store. It returns matching sessions and identifiable excerpts from the original conversations.
post-cutoff
What THIS TEAM has adopted or ruled on that postdates your training cutoff — a small, human-curated set, not a watchlist. Read BEFORE choosing an approach, picking a model, designing a classifier/router/gate, or evaluating a vendor. Also the procedure for proposing and retiring entries. Individual briefs ship as their…