aaaaa
ReviewAudited by ClawScan on May 1, 2026.
Overview
This instruction-only Gmail reply skill is transparent about drafting and optionally sending emails, but users should be careful with Gmail credentials and auto-send settings.
Before installing, confirm you are comfortable giving the agent Gmail access for this client. Prefer approval-before-send mode, use narrowly scoped OAuth where possible instead of broad app passwords, store credentials outside the skill, and enable auto-send only with clear sender, label, and template limits.
Findings (2)
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.
If configured, the agent may be able to access Gmail and send messages as the client.
The skill may use delegated Gmail credentials for the client's account. This is expected for a Gmail auto-reply tool and the artifact says not to store raw passwords in the skill, but users should recognize the account-access implications.
Gmail access: OAuth2 or app password for the client’s Gmail
Use the narrowest available Gmail access, store credentials only in a secure credential manager, and revoke access when no longer needed.
Mistaken configuration could cause emails to be sent automatically under the client's identity.
The skill can send outbound email through a configured send function. The artifact includes approval-first guidance and limits auto-send to explicit configuration, so this is purpose-aligned but still worth user attention.
Only auto-send if the user has clearly configured “auto-send” and you have applied the client’s rules and filters
Keep draft-review mode enabled by default; if using auto-send, restrict it to specific labels, senders, or low-risk templates and periodically audit sent messages.