infrastructure-enumeration

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the agent to capture and output plaintext secrets (e.g., SNMP community strings, IPMI hashes, credentials found) — including calling add_credential(secret="...") and "list all" credentials in summaries — which requires the LLM to handle and emit secret values verbatim.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill explicitly instructs techniques for harvesting sensitive data (DNS zone transfers, TFTP/NFS file grabs, SNMP/LDAP/S MTP enumeration), capturing/storing credentials, and checking/exploiting RCE and persistence vectors (Net-SNMP extend, IPMI cipher 0, writable NFS -> SUID planting), so it contains multiple offensive capabilities that can be directly abused for data exfiltration, credential theft, remote code execution and persistence.

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

• Third-party content exposure detected (high risk: 0.85). Required runtime workflow runs multiple network enumeration commands (e.g., nmap --script dns-zone-transfer, dig axfr, dnsrecon, ldapsearch, snmpwalk, tftp get) that ingest outsider-authored remote service output (from the target host / public network) as readable text into the agent context and evidence files.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (low risk: 0.30). The skill is mainly passive network enumeration and logging, but it explicitly suggests performing a sudo NFS mount (which modifies the local system state) and saving evidence files, so it poses a low but non-zero risk of changing the machine state.