facebook-page-finder
Warn
Audited by Gen Agent Trust Hub on Apr 8, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script explicitly disables global Node.js TLS validation by setting
process.env.NODE_TLS_REJECT_UNAUTHORIZED = '0'. This configuration instructs the environment to ignore certificate errors for all outgoing HTTPS requests, including those to the Apify API, making the connection vulnerable to interception or modification by unauthorized third parties. - [DATA_EXFILTRATION]: The PostgreSQL client configuration in
scripts/facebook-page-finder.jsusesssl: { rejectUnauthorized: false }. This disables certificate verification for the database connection, allowing an attacker positioned between the agent and the database server to intercept sensitive business lead data and database credentials. - [EXTERNAL_DOWNLOADS]: The skill interacts with the Apify platform to perform Google searches via the
apify-clientlibrary. While Apify is a well-known service, the interaction is rendered insecure by the aforementioned global TLS bypass. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8):
- Ingestion points: Data is ingested from the
business_leadsdatabase table and external search results from the Apify Google Search actor (scripts/facebook-page-finder.js). - Boundary markers: None identified; external data is processed and used to update the database without delimiters or instructions to ignore embedded content.
- Capability inventory: The skill has database write access (
UPDATE business_leads) and performs external network requests to Apify. - Sanitization: There is minimal sanitization beyond filtering out generic Facebook keywords (e.g., 'login', 'policies'). This allows potentially untrusted data from search results to be persisted in the internal database.
Audit Metadata