ssrf

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly shows retrieving AWS/GCP/Azure metadata that returns access keys/tokens and then instructs exporting and using those credentials in shell commands (e.g., export AWS_SECRET_ACCESS_KEY=...), which requires embedding secret values verbatim in outputs/commands.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 1.00). High risk — the list contains attacker-controlled OOB domains (interactsh/oast.fun, attacker.com, evil.com), cloud metadata endpoints (169.254.169.254, metadata.google.internal, Azure/Alibaba/Oracle endpoints), Kubernetes/internal service addresses and many IP-encoding/URL-obfuscation variants used in SSRF to exfiltrate credentials or trigger internal requests (even if they are not direct .exe download links).

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This file is an explicit offensive SSRF exploitation guide that instructs attackers how to exfiltrate cloud credentials, perform remote command/data injection (gopher→Redis, SSH key write), and achieve account/system takeover — clearly malicious and intended for abuse.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.65). SKILL.md describes runtime testing that sends attacker-controlled URLs (e.g., interactsh OAST domains) to target URL/webhook/PDF/image/Screenshot import parameters, and the workflow then instructs the agent to “check interactsh log” for resulting DNS/HTTP hits—this OOB log content is outsider-authored free text fetched from a third-party service at runtime.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly instructs to "git clone https://github.com/swisskyrepo/SSRFmap" and then run python3 ssrfmap.py, which fetches remote code from that repository and executes it at runtime (https://github.com/swisskyrepo/SSRFmap).