recon-landscaping

Fail

Audited by Gen Agent Trust Hub on Jul 9, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONCREDENTIALS_UNSAFE
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill employs a pattern where data retrieved from a network source is piped to a Python interpreter for processing.
  • Evidence: curl -sk "https://crt.sh/?q=%25.$TARGET&output=json" | python3 -c "..." in SKILL.md.
  • Note: Although the Python code is static and used for JSON parsing, this pattern is frequently flagged as a security risk.
  • [EXTERNAL_DOWNLOADS]: The skill fetches data from crt.sh to perform certificate transparency-based subdomain enumeration.
  • Evidence: curl -sk "https://crt.sh/?q=%25.$TARGET&output=json" in SKILL.md.
  • [COMMAND_EXECUTION]: The skill makes extensive use of local shell commands and utilities to perform reconnaissance and data probing.
  • Evidence: Automated use of curl, grep, exiftool, and sed to identify directory listings, check for sensitive files, and analyze image metadata.
  • [CREDENTIALS_UNSAFE]: The skill includes automated checks to identify and display sensitive credentials found in exposed log files.
  • Evidence: Grep patterns targeting DB_PASSWORD, DB_USER, and API_KEY within the /wp-content/debug.log file.
Recommendations
  • HIGH: Downloads and executes remote code from: https://crt.sh/?q=%25.$TARGET&output=json - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Jul 9, 2026, 05:20 PM
Security Audit — agent-trust-hub — recon-landscaping