DoctorClaw Email Digest

Security checks across static analysis, malware telemetry, and agentic risk

Overview

This is a coherent email digest skill that openly needs inbox access and may send approved replies, so the main caution is granting it access to private email and delivery channels.

Before installing, decide which inbox the agent may read, keep the email limit and schedule narrow, use read-only access unless you need sending, and deliver digests only to a private trusted channel. Always review drafted replies before approving any send action.

Static analysis

No static analysis findings were reported for this release.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Risk analysis

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

The agent could read private unread emails and, if send access is granted, send replies from the user's account after approval.

Why it was flagged

The skill needs access to a user's email account and may request send permission; this is sensitive account authority, but it is directly tied to the stated digest and reply workflow.

Skill content
**Email access** — Gmail (via Gmail API/skill) or any email provider your agent can read... Optional — **Send access**
Recommendation

Connect only the intended inbox, grant read-only access unless sending is truly needed, and revoke send permissions when no longer required.

What this means

Approved sends can create real outgoing messages from the user's email account.

Why it was flagged

The skill can perform an external action—sending email—including an 'approve all' path, though the artifact states sending happens only after user approval.

Skill content
When the user approves a draft reply ("send reply 1" or "approve all replies"), send the emails.
Recommendation

Review each draft before sending, be cautious with 'approve all replies,' and confirm the recipients and wording before approval.

What this means

Private email information could appear in the chosen delivery channel.

Why it was flagged

The digest and draft replies may contain email subjects, action items, and message details that could be sent to a configured external chat or file destination.

Skill content
**Delivery channel** — Telegram/Discord for digest delivery, or file output
Recommendation

Use a private, trusted delivery destination and avoid shared Telegram or Discord channels for sensitive inbox summaries.

What this means

If scheduling is enabled, the agent may regularly read unread email and deliver digests at configured times.

Why it was flagged

The skill supports recurring scheduled execution, which means it may access the inbox periodically without a fresh manual request each time.

Skill content
Run it on a schedule (morning + afternoon)... Digest schedule — when to run (default: 8:00 AM and 2:00 PM local)
Recommendation

Enable scheduling only intentionally, set the schedule and email limit explicitly, and disable it if recurring inbox access is not desired.