Guru

ReviewAudited by ClawScan on May 10, 2026.

Overview

This is a disclosed Guru integration through Membrane, but installing it means trusting Membrane/OAuth access and allowing actions that can read, create, update, or delete Guru workspace data.

Install only if you trust Membrane and are comfortable connecting it to your Guru workspace. Use a least-privileged Guru account if possible, verify the CLI package, and require explicit confirmation before the agent updates or deletes Guru cards, folders, or comments.

Findings (5)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

You must trust the Membrane CLI package and whatever version is current when installed.

Why it was flagged

The skill relies on an external globally installed npm CLI using the moving @latest version rather than code bundled with the skill.

Skill content
npm install -g @membranehq/cli@latest
Recommendation

Install only from a trusted npm source, consider pinning a known-good CLI version, and review the Membrane package/repository before use.

What this means

The connected account may allow the agent to access Guru data according to the granted permissions.

Why it was flagged

The integration depends on delegated account authentication and ongoing credential refresh for Membrane/Guru access.

Skill content
Membrane handles authentication and credentials refresh automatically
Recommendation

Review the scopes/account used for Guru, prefer least-privilege access where possible, and know how to revoke the Membrane/Guru connection.

What this means

If invoked with sufficient permissions, the agent could change or delete Guru knowledge-base content.

Why it was flagged

The documented Guru action catalog includes mutation and deletion operations for workspace content.

Skill content
| Update Card | update-card | ... | Delete Card | delete-card | Delete a card by ID |
Recommendation

Require clear user confirmation before create, update, or delete operations, and verify target card/folder IDs before running actions.

What this means

Remote setup instructions could influence agent behavior during connection handling.

Why it was flagged

The skill allows remote connection-state responses to provide procedural instructions to the agent.

Skill content
clientAction.agentInstructions (optional) — instructions for the AI agent on how to proceed programmatically.
Recommendation

Treat returned agent instructions as task-scoped guidance only; do not let them override the user's request or normal safety checks.

What this means

Guru workspace data and authentication workflows may pass through the Membrane integration layer.

Why it was flagged

Guru requests and responses are mediated through Membrane rather than only direct local code in the skill.

Skill content
This skill uses the Membrane CLI to interact with Guru.
Recommendation

Review Membrane's data handling and privacy terms before connecting sensitive Guru workspaces.