sablier-withdraw-open-ended-stream

Pass

Audited by Gen Agent Trust Hub on Jun 21, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill facilitates legitimate interactions with the Sablier protocol, a well-known decentralized finance service. It uses the official Sablier indexer infrastructure (indexer.hyperindex.xyz) for stream discovery and links to official documentation for user reference.\n- [SAFE]: Transaction signing is handled with a security-first approach. The skill utilizes the --browser flag of the cast tool to delegate signing to the user's local browser wallet extension (e.g., MetaMask). This prevents the agent from ever having access to the user's private keys or sensitive credentials.\n- [SAFE]: The workflow incorporates mandatory human-in-the-loop confirmation. Before any transaction is broadcast to the blockchain, the skill presents a detailed, human-readable preview of the batch and requires the user to explicitly confirm by typing YES.\n- [SAFE]: Comprehensive pre-flight checks are implemented to ensure transaction success and safety. The skill validates stream status (checking for VOIDED streams which would cause reverts), calculates required protocol fees, estimates gas usage, and verifies that the wallet has a sufficient native token balance before proceeding.\n- [SAFE]: The execution environment is restricted to standard, reputable CLI utilities (cast, jq, curl, bc). The accompanying helper scripts (filter-withdrawable.sh, max-min-fee.sh) are transparent, well-documented, and focused solely on improving efficiency via Multicall3 batching.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 21, 2026, 06:27 PM
Security Audit — agent-trust-hub — sablier-withdraw-open-ended-stream