security-misconfiguration

Fail

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: CRITICALINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONCREDENTIALS_UNSAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external security targets which could contain malicious instructions designed to influence the agent's behavior. \n
  • Ingestion points: The skill ingests HTTP response headers, directory listings, and error pages from remote targets during scanning operations (SKILL.md, payloads.md, guides/debug-endpoint-discovery-guide.md). \n
  • Boundary markers: No explicit delimiters or instructions are provided to the agent to disregard embedded content in tool outputs. \n
  • Capability inventory: The skill utilizes the Bash tool for complex command execution and network operations via curl, wget, and Nmap (SKILL.md). \n
  • Sanitization: There is no automated sanitization described for target responses before they are processed by the agent. \n- [EXTERNAL_DOWNLOADS]: The skill guides the user or agent to install security auditing software from well-known sources. \n
  • Evidence: Instructions to install tools such as Checkov, ScoutSuite, and Prowler via pip, and CloudSploit via npm (guides/cloud-misconfiguration-audit-guide.md, guides/cloud-misconfiguration-checklist-guide.md). These downloads target official repositories or established well-known services. \n- [COMMAND_EXECUTION]: Provides extensive shell command templates for executing security assessment and auditing tools. \n
  • Evidence: Templates for nmap, nikto, hydra, testssl.sh, and cloud CLI tools (aws, gcloud, az) are provided across multiple files (SKILL.md, payloads.md, guides/*.md). These are part of the skill's primary function for security assessment and defense. \n- [CREDENTIALS_UNSAFE]: Lists common default credentials in brute-force dictionary examples for testing purposes. \n
  • Evidence: Common username and password pairs such as 'admin:admin', 'root:root', and 'postgres:postgres' are included in testing wordlists and guides (payloads.md, guides/default-credential-audit-guide.md). These are provided for auditing remote targets and do not represent hardcoded secrets of the agent or author.
Recommendations
  • Contains 2 malicious URL(s) - DO NOT USE
Audit Metadata
Risk Level
CRITICAL
Analyzed
Sep 19, 2026, 04:36 PM