gmgn-callout-score

Pass

Audited by Gen Agent Trust Hub on Nov 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data retrieved from external sources, including blockchain records and social media profile fields (e.g., token symbols, caller handles, and user-provided 'call thesis' text).
  • Ingestion points: Data is ingested via gmgn-cli commands including callout record, portfolio activity, portfolio stats, and market search.
  • Boundary markers: The skill instructs the agent to wrap caller-controlled strings in specific brackets (「」) to distinguish them from system text.
  • Capability inventory: The skill utilizes gmgn-cli to perform network requests and data retrieval, with the output directed back to the user.
  • Sanitization: The skill contains explicit instructions to sanitize these strings by truncating long values, replacing dangerous Unicode control character categories (e.g., Cc, Cs, Co, Zl, Zp), and deleting markdown metacharacters (e.g., [, ], < , >, |) to prevent markdown injection or UI disruption in the agent's output.
  • [EXTERNAL_DOWNLOADS]: The skill depends on the installation of the gmgn-cli tool. This utility is an official resource provided by the skill's vendor (gmgnai), and the instructions for installation and updates via the Node Package Manager (npm) represent standard functional requirements for the skill's operation.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 14, 2026, 02:48 AM