penetration-tester
PassAudited by VirusTotal on May 4, 2026.
Overview
Type: OpenClaw Skill Name: ah-penetration-tester Version: 1.0.0 The skill bundle defines a 'penetration-tester' persona in SKILL.md that is instructed to perform high-risk security activities, including exploit development, social engineering, lateral movement, and establishing persistence. While the instructions include ethical guardrails (e.g., requiring authorization and scope definition), the inclusion of offensive techniques like 'vishing,' 'tailgating,' and 'bypass access control' represents a significant risk profile for an AI agent. No specific malicious payloads, obfuscation, or unauthorized data exfiltration endpoints were found in the provided files.
Findings (0)
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 connected to real tools or systems, the agent could perform intrusive testing steps that may disrupt services, access sensitive data, or affect third parties unless carefully scoped.
These are high-impact dual-use security testing activities. They are aligned with a penetration-testing skill and the artifact also emphasizes authorization and boundaries, but they could be harmful if executed outside an approved engagement.
Network penetration:\n- Network mapping\n- Vulnerability scanning\n- Service exploitation\n- Privilege escalation\n- Lateral movement\n- Persistence mechanisms\n- Data exfiltration\n- Cover track analysis
Use only with written authorization, defined targets, rules of engagement, and explicit human approval before exploitation, lateral movement, persistence testing, social engineering, or data-exfiltration simulations.
Improper use could target employees, collect sensitive information, or create real-world safety, privacy, legal, or reputational harm.
The skill includes social-engineering and physical-security activities, which can affect people and organizations beyond technical systems. This is coherent with broad red-team work, but it needs explicit authorization and safeguards.
Social engineering:\n- Phishing campaigns\n- Vishing attempts\n- Physical access\n- Pretexting\n- Baiting attacks\n- Tailgating\n- Dumpster diving\n- Employee training
Only conduct social-engineering or physical-security testing under a signed engagement with approved scripts, target lists, safety limits, escalation contacts, and opt-out or emergency procedures.