Community code plugin. Review compatibility and verification before install.
Latest release: v1.3.2Download zip
Capabilities
Compatibility
Security Scan
OpenClaw
Suspicious
high confidencePurpose & Capability
The code, dependencies (@honcho-ai/sdk) and hooks implement exactly what the description promises: an Honcho-backed memory adapter, tools for search/context/session, capture hooks, and a CLI for configuring/uploading workspace content. Requiring interaction with Honcho and workspace files is coherent with a memory integration.
Instruction Scope
The package.json (included as the SKILL.md content) and code show the plugin will read environment variables (HONCHO_API_KEY, HONCHO_BASE_URL, HONCHO_WORKSPACE_ID, HONCHO_TIMEOUT_MS) and, via the CLI, scan the user's workspace and ~/.openclaw to find and upload files. The runtime hooks will send conversation/data to the external Honcho service. These side-effects (reading env, scanning local files, transmitting messages/files externally) are beyond a minimal 'memory adapter' expectation for some users and should be acknowledged explicitly in SKILL.md/config — they are not declared in the skill metadata.
Install Mechanism
The registry entry lists no install spec (instruction-only), which is low risk. However package.json includes a postinstall script ('node install.js') and an install.js file is present in the bundle. If the environment that installs this skill runs npm/pnpm postinstall scripts, install.js could execute. There are no external downloads or URL-based installers in the manifest, and distributed code is present in the package.
Credentials
The skill does not declare required environment variables in metadata, yet the code reads process.env.HONCHO_API_KEY, HONCHO_BASE_URL, HONCHO_WORKSPACE_ID, and HONCHO_TIMEOUT_MS and will warn if API key is missing. The CLI also stores API key and baseUrl into ~/.openclaw/openclaw.json. Requesting an API key and writing local config is reasonable for a memory backend, but the omission from requires.env is an inconsistency and the credential usage/transmission to an external service should be considered sensitive.
Persistence & Privilege
The skill does not request always:true and does not appear to modify other skills or system-wide settings beyond writing its own config (~/.openclaw) when the CLI is used. It registers runtime hooks that run during normal plugin operation and will transmit captured memory to Honcho; autonomous invocation is allowed (platform default), which combined with external network access increases blast radius but is expected for a memory plugin.
What to consider before installing
This plugin appears to be a genuine Honcho memory integration, but pay attention to the following before installing: 1) The code uses HONCHO_API_KEY and other HONCHO_* environment variables even though none are declared in the metadata — supply a dedicated API key only if you trust the Honcho service and want your conversation data stored there. 2) The included CLI will scan your workspace and ~/.openclaw, read files (USER.md, MEMORY.md, memory/, canvas/, etc.), and can upload them to Honcho — review which files it will access and back up any sensitive files. 3) package.json contains a postinstall script and an install.js file; verify whether your installation process runs npm postinstall scripts (and inspect install.js) before proceeding. 4) If you need stronger guarantees, prefer a self-hosted HONCHO_BASE_URL (the code supports local URLs) or avoid installing the plugin. 5) If you are not comfortable with an agent potentially sending conversation content to an external service, do not install or restrict the plugin's use to manual (CLI) actions only. Finally, consider auditing the full install.js and capture hook code in your environment and limit granting access to any shared or high-privilege API key.