Security audit

DoctorClaw Client Follow-Up

Security checks for vulnerabilities and agentic risk

Overview

This skill performs disclosed client follow-up drafting and tracking, with manageable privacy and activation concerns users should configure carefully.

Install only if you are comfortable giving the agent access to client/contact records. Limit integrations to the sources and send channels you actually need, require approval before any outreach is sent, and periodically review or delete local follow-up logs that contain client names, deal details, and communication history.

A.I.G

Vulnerability Patterns

Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (2)

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The skill explicitly stores outreach history in local files (`memory/follow-up-log/YYYY-MM-DD.md`) and may save digests to files, but it does not clearly disclose that client names, contact details, deal status, and communication history may persist on disk. This creates a confidentiality and compliance risk, especially on shared hosts or systems without clear data retention and access controls.

Vague Triggers

Medium

Confidence: 96% confidence
Finding: The trigger phrase example, "Check my follow-ups," is broad enough that it could plausibly occur in normal conversation and unintentionally invoke the skill. Because this skill reads client data, drafts outreach, and can later send messages after approval, accidental activation could expose sensitive pipeline information or kick off unwanted business workflows.

Static analysis

No suspicious patterns detected.