Back to skill
Skillv1.0.0

ClawScan security

STM Assistant · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignMar 7, 2026, 6:12 AM
Verdict
benign
Confidence
medium
Model
gpt-5-mini
Summary
The skill's requirements and runtime instructions are coherent with its stated purpose (automated email outreach via an agent inbox) but it relies on an external tool (mcporter/AgentMail) that must be configured outside the skill and that configuration/credentials are not declared by the skill.
Guidance
This skill appears to do what it says: send emails from an agent inbox using mcporter/AgentMail. Before installing: (1) ensure you trust the mcporter tool and the AgentMail/S MTP provider because they will handle all message content and credentials; (2) confirm where mcporter stores credentials and secure that storage (the skill itself does not declare or manage those credentials); (3) be aware that externally hosted signature images can leak recipient/open-tracking data; (4) enforce the approval rules for high-stakes emails (legal/financial) to avoid accidental disclosure; and (5) test sending to controlled addresses before using on real contacts.

Review Dimensions

Purpose & Capability
okName and description describe sending email on behalf of a human; the SKILL.md only asks for mcporter/AgentMail usage, signatures, Reply-To, CC tracking, and inbox monitoring — all directly related to professional email sending.
Instruction Scope
okAll runtime instructions stay within the email-sending domain: they call mcporter agentmail APIs, require a hosted logo, set Reply-To and CC, and prescribe inbox polling and message labeling. The instructions do not ask the agent to read unrelated files or secrets.
Install Mechanism
okInstruction-only skill with no install spec and no bundled code — nothing is downloaded or written to disk by the skill itself. It only requires the external binary 'mcporter' to be present.
Credentials
noteThe skill declares no required env vars, but the SKILL.md depends on mcporter being configured with AgentMail or SMTP credentials outside the skill. Those credentials are effectively required for operation but are not declared; users should confirm how mcporter stores/reads credentials and ensure they trust that configuration.
Persistence & Privilege
okThe skill is not force-included (always: false) and does not request persistent system privileges or modify other skill configs. Normal autonomous invocation is allowed but not elevated here.