← 返回 Skills 市场
samdickson22

ChatGPT Import

作者 samdickson22 · GitHub ↗ · v1.0.2
cross-platform ✓ 安全检测通过
780
总下载
2
收藏
2
当前安装
3
版本数
在 OpenClaw 中安装
/install chatgpt-import
功能描述
Import and embed ChatGPT conversation history into OpenClaw's memory search database for easy archival and retrieval of past chats.
使用说明 (SKILL.md)

ChatGPT History Import

Import your ChatGPT conversations into OpenClaw so they're searchable via memory search.

Workflow

1. Export from ChatGPT

Follow references/export-guide.md to download your conversations.json.

2. Convert to Markdown

python3 scripts/convert_chatgpt.py \
  --input /path/to/conversations.json \
  --output /path/to/chatgpt-history

Options: --min-messages N to skip trivial conversations (default: 2).

3. Embed into SQLite

export OPENAI_API_KEY=sk-...
python3 scripts/bulk_embed.py \
  --history-dir /path/to/chatgpt-history \
  --db /path/to/chatgpt-memory.sqlite

Options: --model, --batch-size, --max-workers, --chunk-size, --api-key.

4. Configure OpenClaw

Add as an extra search path in your OpenClaw config:

memorySearch:
  extraPaths:
    - /path/to/chatgpt-memory.sqlite

Then restart the gateway:

openclaw gateway restart

Important Notes

  • OpenAI API key required. The embed script sends conversation text to api.openai.com for embedding. If your conversations contain secrets, consider filtering them out first or using a scoped API key.
  • No key material stored. The generated DB does not store your API key.
  • Back up first. The embed script will refuse to overwrite an existing output DB.
  • Embeddings cost money — but it's cheap. ~2,400 conversations cost ~$0.15 with text-embedding-3-small.
安全使用建议
This skill appears to do exactly what it claims: convert your ChatGPT export into markdown and embed it using OpenAI embeddings into a local SQLite DB for OpenClaw. Before running it, consider the following: - The scripts send your conversation text to OpenAI's embeddings endpoint (api.openai.com). If conversations include sensitive secrets, remove or scrub them first, or use a scoped API key. The SKILL.md notes this. - The registry metadata does not list the required OPENAI_API_KEY env var — supply a key via OPENAI_API_KEY or --api-key when running. Treat this as a documentation gap rather than malicious behavior, but be cautious. - Review and run the scripts locally in a controlled environment. The code is included and readable; verify paths and back up your data before running (the script refuses to overwrite an existing DB but you should still back up exports). - If you need stronger privacy, consider using a provider or on-prem embedding solution instead of sending transcripts to the public OpenAI API. If you want, I can highlight the exact lines where the API call occurs and where the key is read, or produce a checklist to run this safely (e.g., run in a temp directory, use a scoped key, inspect outputs).
功能分析
Type: OpenClaw Skill Name: chatgpt-import Version: 1.0.2 The OpenClaw AgentSkills bundle is designed to import ChatGPT conversation history into OpenClaw's memory search. All files, including `SKILL.md`, `scripts/convert_chatgpt.py`, and `scripts/bulk_embed.py`, align with this stated purpose. The `SKILL.md` provides clear user instructions without any prompt injection attempts against the agent. The Python scripts perform necessary file I/O and make legitimate API calls to `api.openai.com` for embeddings, transparently handling the OpenAI API key from environment variables or arguments. There is no evidence of data exfiltration to unauthorized endpoints, persistence mechanisms, or other malicious activities.
能力评估
ℹ Purpose & Capability
Name/description, SKILL.md, and the included scripts all align: converting ChatGPT's conversations.json to markdown and embedding those texts into an OpenClaw-compatible SQLite DB. However, the registry metadata lists no required environment variables or primary credential, while the runtime requires an OpenAI API key (OPENAI_API_KEY or --api-key). This is a metadata/documentation inconsistency (not functional maliciousness).
✓ Instruction Scope
SKILL.md precisely instructs export → convert → embed → register DB with OpenClaw. The scripts only read user-provided export files and the output directory, call OpenAI's embeddings endpoint (api.openai.com), and write a local SQLite DB. There are no instructions to read unrelated system files or send data to unexpected endpoints. The README explicitly warns about sending secrets and about backups.
✓ Install Mechanism
No install spec; this is an instruction+script bundle with no network install step. All code is included in the package and uses standard Python stdlib; nothing is downloaded or extracted at install time.
ℹ Credentials
The functionality legitimately requires an OpenAI API key to call the embeddings API, and filesystem access to the user's exported conversations and to write the output DB. The metadata, however, declares no required env vars or primary credential — this mismatch should be corrected. No other credentials or unrelated environment access are requested.
✓ Persistence & Privilege
The skill is not always-enabled and does not request elevated or permanent presence. It does not modify other skills or system-wide config by itself; it only creates a local DB file at a user-specified path and suggests user actions to add that DB to OpenClaw's search paths.
如何使用
  1. 确保已安装 OpenClaw（本地或 Docker 部署）
  2. 在对话框中输入安装命令：/install chatgpt-import
  3. 安装完成后，直接呼叫该 Skill 的名称或使用 /chatgpt-import 触发
  4. 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
v1.0.2
- Added "Important Notes" section detailing security and cost considerations for embedding conversations. - Clarified that an OpenAI API key is required and shared advice on handling secrets and API key usage. - Noted that the output database does not store your API key. - Reminded users to back up prior to embedding; script refuses to overwrite existing databases. - Provided an example cost for embeddings to help users estimate expenses.
v1.0.1
- Documentation updated for clarity in the "Configure OpenClaw" step. - Now explicitly instructs users to add as an extra search path, removing the option to replace the main memory DB. - No functional or code changes in this release.
v1.0.0
Initial release - converts ChatGPT conversations.json to markdown, bulk-embeds into OpenClaw SQLite, configures memory search.
元数据
Slug chatgpt-import
版本 1.0.2
许可证 —
累计安装 2
当前安装数 2
历史版本数 3
常见问题

ChatGPT Import 是什么？

Import and embed ChatGPT conversation history into OpenClaw's memory search database for easy archival and retrieval of past chats. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 780 次。

如何安装 ChatGPT Import？

在 OpenClaw 或 Claude Code 对话框中运行命令「/install chatgpt-import」即可一键安装，无需额外配置。

ChatGPT Import 是免费的吗？

是的，ChatGPT Import 完全免费（开源免费），可自由下载、安装和使用。

ChatGPT Import 支持哪些平台？

ChatGPT Import 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 ChatGPT Import？

由 samdickson22（@samdickson22）开发并维护，当前版本 v1.0.2。

推荐 Skills
self-improving agent
Captures learnings, errors, and corrections to enable continuous improvement. Use when: (1) A command or operation fails unexpectedly, (2) User corrects Clau...
⬇ 463283 · by pskoett
Skill Vetter
Security-first skill vetting for AI agents. Use before installing any skill from ClawdHub, GitHub, or other sources. Checks for red flags, permission scope, and suspicious patterns.
⬇ 259410 · by spclaudehome
Polymarket
Query Polymarket prediction markets. Check odds, find trending markets, search events, track price movements.
⬇ 232503 · by joelchance
Self-Improving + Proactive Agent
Self-reflection + Self-criticism + Self-learning + Self-organizing memory. Agent evaluates its own work, catches mistakes, and improves permanently. Use when...
⬇ 200423 · by ivangdavila
Github
Interact with GitHub using the `gh` CLI. Use `gh issue`, `gh pr`, `gh run`, and `gh api` for issues, PRs, CI runs, and advanced queries.
⬇ 191113 · by steipete
ontology
Typed knowledge graph for structured agent memory and composable skills. Use when creating/querying entities (Person, Project, Task, Event, Document), linkin...
⬇ 190156 · by oswalpalash

💬 留言讨论