redbook-analytics
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is entirely instructional and does not contain any executable code, scripts, or automated network operations. All suggested actions are for the user to perform or for the agent to use in its logical reasoning.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data provided by the user, such as account metrics and post details. While this presents a theoretical attack surface, it is considered safe due to the lack of exploitable capabilities.\n
- Ingestion points: Data is collected in 'Step 1: Information Collection' through user-provided text prompts.\n
- Boundary markers: The instructions do not specify any delimiters (such as XML tags or special characters) or warnings to ignore instructions that might be embedded in the user's data.\n
- Capability inventory: Across all instructions, there are no references to dangerous tools or capabilities. The skill does not perform file system operations, network requests (curl, wget), or dynamic code execution (exec, eval).\n
- Sanitization: There is no mention of sanitizing, escaping, or validating the external content provided by the user.
Audit Metadata