Webhook Router

• Analyzed: May 12, 2026, 4:19 AM
• Findings: 1
• Version: 1.0.0
• Hash: 25d47b2272d01b874a4b6e46546dfccc211eb482daaef698e13a5e0f5590611e
• Source: palm
• External report: View on VirusTotal

Findings (1)

This skill is classified as suspicious due to multiple critical vulnerabilities related to insecure handling of untrusted input. The `register.sh` script is vulnerable to shell injection via `sed` when creating handler templates, potentially allowing arbitrary file modification or command execution if an attacker controls the `source-type` argument. Furthermore, all handler scripts (`handlers/custom.sh`, `handlers/generic.sh`, `handlers/github.sh`) pass unsanitized data derived from untrusted webhook payloads directly to `message send` and `vault write` commands. Given that OpenClaw instructions are executed by an AI agent, this presents a direct prompt injection vulnerability against the agent, allowing an attacker to inject arbitrary instructions or commands via crafted webhook payloads. A hardcoded `X-Hook-Token` in `SKILL.md` and `register.sh` is also a minor security flaw.