Lattice
ReviewAudited by ClawScan on May 10, 2026.
Overview
This Lattice skill needs review because it gives broad, weakly bounded authority over HR data and lists unrelated sensitive financial, medical, and purchasing domains.
Review this skill carefully before installing. If used, grant only least-privilege Lattice/Membrane access, require explicit confirmation before any data-changing action, and avoid using it for financial, medical, purchasing, or other non-Lattice tasks unless the provider documents and scopes those capabilities clearly.
Findings (3)
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.
An agent could treat the skill as authority to modify important HR or business records, or to handle purchasing/payment-style tasks, without enough built-in guardrails.
The instruction text uses broad mutation-oriented language and includes purchasing/payment-related objects, but the visible artifact does not provide clear limits or confirmation requirements for high-impact actions.
“Manage Persons, Organizations, Roles, Activities, Notes, Files” ... “Purchase Request” ... “Ticket Purchase” ... “Order” ... “Payment”
Require explicit user confirmation for any create, update, delete, role/person change, file operation, purchase, or payment-related action, and document which Lattice endpoints are actually supported.
Users may not understand what account or token permissions the skill will use to access Lattice or Membrane data.
The registry requirements say no credential is declared, while capability signals indicate OAuth or sensitive credentials are needed; this makes the delegated account boundary less clear to users.
“Primary credential: none” ... “Required env vars: none” ... Capability signals: “requires-oauth-token”, “requires-sensitive-credentials”
Before installing, verify the OAuth/account scopes, use a least-privilege account, and avoid granting admin-level access unless required and documented.
The agent may over-trust the skill as a general-purpose handler for sensitive financial, medical, or other non-Lattice data.
The skill is framed as a Lattice performance-management integration, but its visible overview lists many unrelated and highly sensitive domains, which can mislead an agent or user about the intended scope.
“Lattice is a performance management platform” ... “Medical Record” ... “Bank Account” ... “Credit Card” ... “Cryptocurrency” ... “National Security”
Narrow the SKILL.md content to actual Lattice resources and remove unrelated sensitive domains unless each is genuinely supported, justified, and protected by explicit user controls.