Community update ( OpenFang v0.5.7 )
Before I get into the update, I want to briefly address something.
Over the past days I haven’t been as active, mainly due to a situation involving a holder controlling more than 7% of the supply, which created unnecessary pressure on the distribution.
Fortunately, that position has now been fully exited.
To anyone currently holding more than 4%, I kindly ask you to consider reducing your position. Large concentrations like this can discourage new participants and create hesitation for incoming holders.
I genuinely ask this with respect, let’s protect the distribution and keep things healthy for the long term.
Now, onto the actual update.
After ~12 days, @openfangg just shipped a major release, and this one goes deeper on security, infrastructure, and real-world usability.
v0.5.7 is a meaningful step forward:
• stronger security foundations
• more reliable agent execution
• improved developer experience
• growing ecosystem momentum
Security & Core Improvements
• Upgraded to Argon2id password hashing (replacing SHA-256) with random salts
• Added SSRF allowlist control for safer internal service access
• Completed a full CSO security audit
• Improved credential handling and execution policies
Security continues to be treated as a first class priority.
New Capabilities
• Image generation pipeline (DALL-E / GPT-Image)
• SearXNG integration for self-hosted search
• Docker sandbox runtimes for safer execution
• Shell runtime support for agent-level operations
• WeCom channel adapter added
These push OpenFang further toward production grade usage.
Agent & Model Enhancements
• Expanded embedding auto-detection across multiple providers
• Improved tool call recovery for Llama-family models
• Increased Ollama context window (128K / 16K output)
• Smarter fallback handling and model routing
This directly improves reliability in complex workflows.
Performance & Stability
• Fixed multiple streaming and context overflow issues
• Increased agent execution timeout (120s → 600s)
• Reduced unnecessary loops and credit waste
• Improved overall system stability
Developer & UX Improvements
• Skill hot-reload + Git-based installs
• Dynamic provider/model loading (43 providers supported)
• Improved dashboard + real time streaming
• Better debugging and error visibility
Ecosystem Progress
• 59 PRs merged total (9 in this release)
• 33 bugs fixed
• 2,186+ tests passing
• 0 clippy warnings
Clear signal of accelerating development and community contribution.
https://t.co/rJedvJ9yc0
gFang @openfangg@Akashi203 v0.5.7 dropped with no Windows binaries — installer fails and cargo build
is broken too (rmcp dependency).
Might just need a CI re-run
v0.5.6 works as a workaround for now
Filed https://t.co/wReWs1qBof
Day 4 with @openfangg
I just love it , I don't want to spoil too much before its live but I honestly love how good OpenFang is at making things
thats it.
@Akashi203@openfangg Great work Jaber, will hopefully have time to start building on Openfang this weekend.
Been reading a couple of great reviews in the $Openfang community, looking forward to getting my fangs stuck in 🐍
@openfangg Came across openfang the other day, absolutely amazing tech and work rate. Then came across the community page, refreshing to see a community page where they put the tech/project before their own greed.
Keep up the great work 🐍🐍🐍