firebase-supabase-attack

Pass

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses local shell commands, including curl and grep, to perform network reconnaissance and extract configuration details from downloaded files. It also uses python3 -c to dynamically parse JSON responses from remote APIs.- [EXTERNAL_DOWNLOADS]: The skill fetches external content such as JavaScript files from target websites and interacts with Firebase and Supabase REST APIs. These domains are well-known service providers.- [DATA_EXFILTRATION]: The skill is designed to retrieve data from misconfigured cloud databases and save the results to the local /tmp directory for analysis.- [PROMPT_INJECTION]: The skill demonstrates a surface for indirect prompt injection by processing untrusted data from target websites (HTML, JS) and API responses. Ingestion points: curl outputs in Phases 1, 2, 3, 5, 6, and 8. Boundary markers: Absent. Capability inventory: Subprocess execution (curl, python3), file system writes (/tmp), and network operations. Sanitization: The skill lack explicit sanitization or escaping of the content retrieved from external targets before it is processed or displayed to the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 31, 2026, 04:47 PM
Security Audit — agent-trust-hub — firebase-supabase-attack