saas-launch-audit

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the Bash tool to execute local shell scripts (scripts/tech-audit.sh and scripts/competitor-visibility.sh) to perform technical audits. These scripts use system utilities like curl and openssl to check headers and SSL status. URL arguments are quoted to mitigate basic injection.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted content from external URLs. Ingestion points: Data is fetched from the target URL using WebFetch and browser tools. Boundary markers: The instructions do not define delimiters or ignore-instructions for the external content. Capability inventory: The skill can execute local scripts via Bash and perform web searches. Sanitization: While the audit script extracts the domain for technical checks, the natural language content retrieved is not sanitized before LLM processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 03:57 AM