superfluid

Pass

Audited by Gen Agent Trust Hub on Mar 3, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill contains several utility scripts (abi.mjs, balance.mjs, metadata.mjs, tokenlist.mjs) designed to be executed via bunx. These scripts dynamically load official Superfluid libraries from the NPM registry, such as @sfpro/sdk, @superfluid-finance/tokenlist, and @superfluid-finance/metadata. These downloads are intended for the skill's core functionality of resolving protocol-specific data.
  • [DATA_EXFILTRATION]: Real-time balance and metadata resolution logic involves network requests to protocol-associated domains, including superapi.kazpi.com, superfluid.dev, and superfluid.pro. These requests are used to retrieve public on-chain information and do not target sensitive local files or user credentials.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a data ingestion surface as it processes responses from external APIs and subgraphs. 1. Ingestion points: balance.mjs and metadata.mjs. 2. Boundary markers: Not present in script outputs (raw JSON). 3. Capability inventory: No file-system writes or arbitrary command execution capabilities identified. 4. Sanitization: Not explicitly implemented in the data retrieval scripts. The risk is minimized as the data sources are restricted to official protocol infrastructure.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 3, 2026, 07:09 AM