Skip to content
Forge
AI Agents & Assistants
A reliability layer that adds guardrails (retry handling, response validation, tool‑call parsing) to self‑hosted LLM tool‑calling workflows.
Python
·
Latest v0.8.1 · 17d ago
Security brief →
Features
-
Adds reliable guardrails (rescue parsing, retry nudges, response validation) with zero required steps
-
Three usage modes: drop‑in proxy server, full WorkflowRunner with lifecycle management, and reusable middleware for custom orchestrators
-
Supports multiple LLM backends including Ollama, llama-server, vLLM, Anthropic API
Review required
v0.8.1
Bug fix
·
Auth
RCE / SSRF
Rescued tool calls + error handling
Config change
v0.8.0
Breaking risk
·
Auth
Auth handling breaking change
No immediate action
v0.7.6
Bug fix
·
Capture inline reasoning
Config change
v0.7.5
Breaking risk
·
Bounded reasoning replay + Anthropic enhancements
No immediate action
v0.7.4
Breaking risk
·
Tool-error correction + max‑errors flag
Weekly OSS security release digest.
The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.
No spam, unsubscribe anytime.
About
Languages
Python
·
TypeScript
·
Dockerfile
Search tools, categories, lists, and users
Use ↑↓ to navigate, Enter to open, Esc to close
No results for ""
⌘K to open
↑↓ navigate
⏎ open
