BlockBeats News, March 3 — The open-source AI Agent framework OpenClaw released version v2026.3.2, which includes multiple new features, security enhancements, and over 150 bug fixes, contributed by 93 developers.
Major new features in this release include a native PDF analysis tool supporting Anthropic and Google as PDF processing backends, with configurable extraction fallback strategies and page/size limits; the SecretRef credential referencing mechanism extended to 64 targets, covering runtime collectors, planning/execution/audit workflows, with unresolved references immediately causing errors on active interfaces; a new STT (Speech-to-Text) API supporting transcription of audio files via configured service providers; Telegram message streaming default switched to “partial” mode for real-time preview; and a new MiniMax-M2.5-highspeed model added to the provider directory.
This version introduces four breaking changes: the default tool configuration for new installations switched from a broad programming toolset to “messaging”; ACP scheduling enabled by default; the plugin HTTP routing registration API changed from registerHttpHandler to explicitly declared registerHttpRoute requiring authentication; Zalo Personal no longer depends on external CLI binaries, now running purely on JS runtime.
Security fixes include reinforcement of Gateway loopback WebSocket, prevention of duplicate plugin route registration, pre-parse Webhook authentication, and protection against symbol link escape in skill workspaces.
Disclaimer: The information on this page may come from third parties and does not represent the views or opinions of Gate. The content displayed on this page is for reference only and does not constitute any financial, investment, or legal advice. Gate does not guarantee the accuracy or completeness of the information and shall not be liable for any losses arising from the use of this information. Virtual asset investments carry high risks and are subject to significant price volatility. You may lose all of your invested principal. Please fully understand the relevant risks and make prudent decisions based on your own financial situation and risk tolerance. For details, please refer to
Disclaimer.
Related Articles
SBI Ripple Asia Completes XRPL Token Issuance Platform for Business Integrations
SBI Ripple Asia has developed a token issuance platform on the XRP Ledger, allowing businesses to integrate token management via API without overhauling customer interfaces. This approach encourages blockchain adoption by blending seamlessly with existing infrastructure.
CryptoNewsFlash16m ago
The Ethereum Foundation teamed up with Biconomy to launch ERC-8211: AI agents can handle multiple DeFi contract interactions in a single transaction
The Ethereum Foundation and Biconomy have teamed up to launch the ERC-8211 standard, enabling AI agents to dynamically execute multi-step DeFi actions within a single signed transaction, determining parameters in real time based on on-chain state and overcoming the static limitations of ERC-4337. This standard does not require a fork at the protocol layer, making it easier for developers to adopt and further accelerating the integration of AI and DeFi automation.
動區BlockTempo1h ago
Chaos Labs Exits Aave Role as Governance Dispute Deepens Ahead of V4
Chaos Labs is leaving Aave due to governance disagreements, marking a significant shift in the protocol's risk management approach. Founder Omer Goldberg cited unprofitability, recent departures of other major contributors, and fundamental misalignment on risk handling as reasons for the split.
CryptoNewsFlash1h ago
XRP Whales Accumulate as Ripple Executives Head to Key XRPL Event in Japan
XRP has gained 5% from a support level of $1.28 due to increased whale accumulation, with large holders now buying over 11 million XRP daily. Ripple executives are attending a significant XRPL event in Japan, though broader market sentiment remains cautious amid geopolitical tensions.
CryptoNewsFlash2h ago
Web3 Security Provider Certik Opens AI Auditing Tool Access to Global Developers
Certik has transitioned its AI auditor to a public-facing tool, achieving an 88.6% hit rate in identifying security vulnerabilities. This proactive system reduces false positives and integrates real-time threat intelligence, enhancing security in Web3 development environments.
Coinpedia4h ago
