blitz-list-builder
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The
find-contacts.tsscript performs network requests toapi.useblitz.comto retrieve company and employee information. While this is the intended purpose of the tool, it represents communication with a third-party service that is not part of the trusted vendor or organization lists. - [INDIRECT_PROMPT_INJECTION]: The skill acts as a data bridge where information from the Blitz API (such as job titles) is intended to be evaluated by a subsequent AI qualification step, such as
/icp-prompt-builder. This creates a potential surface for indirect prompt injection, as an attacker who controls the data returned by the API (e.g., via a LinkedIn profile) could include malicious instructions designed to manipulate the behavior of downstream AI models. - Ingestion points: Data is ingested from the Blitz API (
/api/enrichment/company) inscripts/find-contacts.tsand from user-provided domain files. - Boundary markers: The script produces a CSV output but does not include explicit delimiters or instructions to ignore embedded AI commands in the resulting fields, which could be interpreted by a model reading the CSV.
- Capability inventory: The skill possesses capabilities for reading and writing local files and making outbound network connections to external APIs.
- Sanitization: The script does not sanitize the employee data fields returned by the API (like titles, names, and industry information) before including them in the output CSV file.
Audit Metadata