← 返回 Skills 市场
Chroma Memory
作者
iPythoning
· GitHub ↗
· v1.0.1
· MIT-0
153
总下载
0
收藏
2
当前安装
2
版本数
在 OpenClaw 中安装
/install chroma-memory
功能描述
Stores and semantically retrieves per-turn customer conversation data with auto-tagging and customer isolation using ChromaDB.
使用说明 (SKILL.md)
chroma-memory — Per-Turn Conversation Memory with ChromaDB
Long-term vector memory for customer conversations. Stores every turn with customer isolation, auto-tags quotes and commitments, enables semantic retrieval across sessions.
Commands
| Command | Description |
|---|---|
chroma:store |
Store a conversation turn (auto-called after each turn) |
chroma:search \x3Cquery> |
Semantic search across conversation history |
chroma:recall \x3Ccustomer_id> |
Recall recent history for a returning customer |
chroma:snapshot |
Store daily CRM snapshot as fallback (L4) |
chroma:stats |
Show storage statistics |
Usage
# Store a turn (normally auto-triggered by hook)
chroma:store --customer "+971501234567" --turn 5 --user "What's the price for 500 units?" --agent "Let me prepare a detailed quote..." --stage qualifying --topic pricing
# Search history
chroma:search "Dubai customer pricing discussion" --customer "+971501234567" --limit 5
# Recall for returning customer (auto-triggered when gap > 7 days)
chroma:recall "+971501234567" --limit 10
# Daily CRM snapshot (triggered by HEARTBEAT #12)
chroma:snapshot
# Stats
chroma:stats
Architecture
This skill implements Layer 3 (L3) and Layer 4 (L4) of the 4-layer Anti-Amnesia system:
- L3: Every conversation turn → ChromaDB with customer_id isolation + auto-tagging
- L4: Daily CRM snapshot → ChromaDB as disaster recovery fallback
Auto-Tagging
Turns are automatically tagged based on content analysis:
| Tag | Trigger |
|---|---|
has_quote |
Price/cost/quote discussed |
has_commitment |
Promise made by either party |
has_objection |
Customer objection detected |
has_order |
Order/purchase confirmed |
has_sample |
Sample request discussed |
Customer Isolation
All data is partitioned by customer_id (phone number). Queries always include where={"customer_id": ...} to ensure strict tenant isolation.
Dependencies
chromadbskill (vector database, install via ClawHub)- OpenClaw Gateway with session-memory hook enabled
安全使用建议
The core issue is mismatch: the skill promises ChromaDB-backed semantic vector memory but the bundled code implements only local JSON storage and lexical ranking. Before installing, decide whether you want a local file-based memory (this code) or true ChromaDB/vector embeddings. If you accept local storage, confirm where files will be written (OPENCLAW_HOME or HOME), whether data-at-rest is encrypted, retention/cleanup policies, and whether phone numbers/PII are acceptable to store on disk. If you expect a ChromaDB integration, ask the publisher to provide the real connector or remove the misleading claims. Also note the script requires Node.js to run (not declared) — inspect and test the code in a sandbox with non-production data first.
能力标签
能力评估
Purpose & Capability
SKILL.md and description claim ChromaDB integration and semantic (vector) retrieval; the provided chroma.mjs contains no network/DB calls, no use of chromadb or embeddings — it stores JSON files under ~/.openclaw/memory/chroma and uses simple lexical overlap + heuristics for search. The declared dependency on a 'chromadb' skill and the ChromaDB branding are therefore misleading.
Instruction Scope
Runtime instructions describe 'auto-called' hooks, tenant isolation via where filters, and semantic searches. The code is a standalone Node CLI that reads/writes local JSON files and does not implement semantic vector search or any hook integration. SKILL.md references HEARTBEAT triggers and OpenClaw Gateway behavior that are not visible in the code.
Install Mechanism
There is no install spec (instruction-only), which is low risk. However, the package includes a Node .mjs script but the skill metadata lists no required binaries; 'node' is effectively required to run chroma.mjs but isn't declared. Nothing in the install path downloads remote code or runs network installers.
Credentials
The skill declares no required secrets or env vars. The code reads OPENCLAW_HOME or HOME and optionally CRM_SNAPSHOT_DATA (used for snapshot piping) — these env vars are not documented in requires.env. None are sensitive credentials, but CRM_SNAPSHOT_DATA could contain bulk CRM data and is not flagged as a required/optional env var in the metadata.
Persistence & Privilege
always:false and normal invocation. The code writes persistent files into the user's HOME (or OPENCLAW_HOME) under a .openclaw/memory/chroma path and will create directories and many JSON files. It does not modify other skills or system-wide settings, but it does persist potentially sensitive conversation data on disk.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install chroma-memory - 安装完成后，直接呼叫该 Skill 的名称或使用
/chroma-memory触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.1
chroma-memory 1.0.1
- Documentation updates only: SKILL.md reformatted, with no changes to commands, features, or usage.
- No functional changes to the core skill logic.
v1.0.0
Initial publish from b2b-sdr-agent-template v3.3.1
元数据
常见问题
Chroma Memory 是什么？
Stores and semantically retrieves per-turn customer conversation data with auto-tagging and customer isolation using ChromaDB. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 153 次。
如何安装 Chroma Memory？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install chroma-memory」即可一键安装，无需额外配置。
Chroma Memory 是免费的吗？
是的，Chroma Memory 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Chroma Memory 支持哪些平台？
Chroma Memory 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Chroma Memory？
由 iPythoning（@ipythoning）开发并维护，当前版本 v1.0.1。
推荐 Skills