← 返回 Skills 市场

Hippocampus Openclaw Onboarding

Name: Hippocampus Openclaw Onboarding
Author: cezexpl

作者 cezexPL · GitHub ↗ · v0.1.0 · MIT-0

cross-platform ⚠ suspicious

359

总下载

当前安装

版本数

在 OpenClaw 中安装

/install hippocampus-openclaw-onboarding

功能描述

Bootstrap OpenClaw with Hippocampus memory under a branded, repeatable setup: workspace, agent ID, API key or bootstrap token, and MCP wiring.

使用说明 (SKILL.md)

Hippocampus OpenClaw Onboarding

Use this skill when connecting a new OpenClaw instance or workspace to Hippocampus memory.

Setup Goals

define a stable workspace name
assign a durable root agent_id
configure gateway URL and authentication
ensure sub-agents inherit scoped memory identity

Preferred Flow

Register or sign in through the Hippocampus portal.
Create a root OpenClaw agent in the dashboard.
Copy the generated bootstrap one-liner.
Run: npx hipokamp-mcp setup --bootstrap-token \x3Ctoken> --gateway \x3Cgateway-origin>
Let hipokamp-mcp write local config under ~/.hipokamp/config.json.
Verify health and first store/search round trip.

Guidance

Use one root agent per OpenClaw instance.
Keep sub-agent IDs under the same workspace namespace.
Prefer bootstrap-first setup over pasting a long-lived API key.
Do not reuse credentials across unrelated workspaces.

hippocampus-memory-core once setup is complete
hippocampus-subagent-memory for child-agent isolation
@hippocampus/openclaw-context-engine for native lifecycle integration

安全使用建议

This skill is plausible but has red flags you should resolve before running anything: (1) verify the correct project/package name and spelling with the Hippocampus vendor — the SKILL.md uses 'hipokamp' (missing 'c'), which could be a typo or a malicious typosquat; (2) avoid blindly running `npx` commands that fetch and execute code from npm — prefer a vetted, pinned package or review the package source before running; (3) treat the bootstrap token as sensitive: do not paste it into untrusted prompts or UIs; ask how the token is stored (inspect ~/.hipokamp/config.json after setup and set strict file permissions); (4) request an install spec or signed release (GitHub release tarball, checksums) if you need to automate this for multiple agents; (5) if you proceed, test in an isolated environment (throwaway account / VM) first to confirm the package behavior and config contents. If you cannot verify the publisher or package repository, do not run the npx command.

功能分析

Type: OpenClaw Skill Name: hippocampus-openclaw-onboarding Version: 0.1.0 The skill bundle provides a standard onboarding workflow for integrating OpenClaw with a service called Hippocampus. It outlines a setup process involving the creation of an agent ID and the execution of a configuration command via `npx hipokamp-mcp` to write local settings to `~/.hipokamp/config.json`. The instructions in SKILL.md are transparent, aligned with the stated purpose of service bootstrapping, and contain no evidence of data exfiltration, credential theft, or malicious prompt injection.

能力评估

ℹ Purpose & Capability

The skill's stated purpose (bootstrapping OpenClaw with Hippocampus memory) aligns with the actions described (register, obtain bootstrap token, run a setup helper, write local config). However the SKILL.md uses inconsistent branding/spelling (Hippocampus vs 'hipokamp' vs '~/.hipokamp' and package name 'hipokamp-mcp'), which could be a benign typo or indicate a typo-squatted/misnamed package. This mismatch is unexpected and worth verifying.

⚠ Instruction Scope

The runtime instructions tell the operator to run `npx hipokamp-mcp setup --bootstrap-token <token> --gateway <gateway-origin>` which will fetch and execute code from npm and write a config file under the user's home directory (~/.hipokamp/config.json). The SKILL.md does not instruct the agent to read unrelated system files, but it does direct execution of remotely fetched code and persistent storage of credentials — actions that go beyond a purely declarative onboarding document and carry risk if the package or token handling is malicious or mistaken.

⚠ Install Mechanism

There is no install spec in the manifest, but the instructions implicitly rely on npx to download and run 'hipokamp-mcp' from the npm registry. Running npx executes remote code at runtime; combined with the inconsistent package/branding spelling, this raises a higher-risk install mechanism (remote execution without a pinned, verifiable source or SHA).

ℹ Credentials

The skill metadata declares no required environment variables, yet the instructions explicitly require a bootstrap token (sensitive) and will write local config that likely contains credentials. The lack of a declared primaryEnv / required envs is an omission: the skill uses sensitive data but doesn't list or justify it in the manifest. Writing credentials to ~/.hipokamp is proportionate for onboarding but should be made explicit and documented (storage format, permissions, rotation guidance).

ℹ Persistence & Privilege

The skill does not request 'always: true' and does not modify other skills, but the onboarding flow writes persistent configuration to the user's home directory (~/.hipokamp/config.json). That persistent write is expected for onboarding but is a point of lasting privilege (stored tokens/config) and should be audited by the user for content and permissions.

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 hippocampus-openclaw-onboarding
安装完成后，直接呼叫该 Skill 的名称或使用 /hippocampus-openclaw-onboarding 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

v0.1.0

Initial release of hippocampus-openclaw-onboarding. - Provides step-by-step onboarding instructions for connecting OpenClaw to Hippocampus memory. - Details necessary setup actions: workspace naming, agent ID assignment, and authentication configuration. - Explains best practices for agent hierarchy and credential management. - Lists related packages for post-setup operations and integration.

元数据

Slug hippocampus-openclaw-onboarding

版本 0.1.0

许可证 MIT-0

累计安装 0

当前安装数 0

历史版本数 1

常见问题