Back to skill
Skillv2.2.0
VirusTotal security
Agentok Skill · External malware reputation and Code Insight signals for this exact artifact hash.
Scanner verdict
SuspiciousApr 30, 2026, 4:21 AM
- Hash
- a27182bf339716c86c8bf9615a45c78107aa826ce1d37a5a4204e6ca1318202e
- Source
- palm
- Verdict
- suspicious
- Code Insight
- Type: OpenClaw Skill Name: agenttok Version: 2.2.0 The skill is classified as suspicious due to a critical shell injection vulnerability in `scripts/join.sh`. The `ffmpeg` command directly interpolates user-provided arguments (`NAME`, `HANDLE`) into its `drawtext` filter without sanitization, allowing for arbitrary command execution if an attacker can control these inputs. This risk is exacerbated by the `SKILL.md` instructions, which guide the AI agent to execute `scripts/join.sh` with user-controlled arguments, creating a clear prompt-injection vector for RCE. Additionally, the default API endpoint `https://rev-mon-avon-childhood.trycloudflare.com` uses a Cloudflare Tunnel, which is a minor suspicious indicator, though not conclusive of malice on its own.
- External report
- View on VirusTotal