bigquery-oulang-analysis

Pass

Audited by Gen Agent Trust Hub on Apr 7, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to use shell execution tools (such as bash or serena:execute_shell_command) to run BigQuery (bq) CLI commands, allowing for direct system interaction.
  • [DATA_EXFILTRATION]: The skill accesses sensitive production data, including user phone numbers and nicknames from the users and listings tables, as well as financial transaction records.
  • [EXTERNAL_DOWNLOADS]: The skill references external URLs for data retrieval and database proxying, including https://posthog.pime.ai/api/projects/1/query/ and cors.trigox.workers.dev/db.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection.
  • Ingestion points: Data is ingested from tables like listings and messages which contain content provided by untrusted external users.
  • Boundary markers: Absent. No delimiters or instructions are provided to the agent to ignore potentially malicious commands embedded in the retrieved data.
  • Capability inventory: The agent is equipped with shell command execution capabilities and database access tools.
  • Sanitization: Absent. The skill does not implement filtering or escaping of user-generated content before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 7, 2026, 07:22 AM