GraphARC: Letting LLMs Autonomously Plan Agent Workflow Topologies at Runtime
Desperate-Ad-9679 · reddit · 2026-08-06
A developer open-sourced GraphARC, an agent graph orchestration framework built on LangGraph. It aims to solve the conflict where loops are easy to ship but impossible to audit, while graphs are auditable but hard to hand-author.
Core Mechanics:
- Dynamic Generation: Uses a local model like Qwen3:8b to dynamically propose a workflow topology at runtime based on a single natural language query.
- Admission Gate: A deterministic gate checks the proposed graph against registry, policy, budget, and acyclicity. Only admitted graphs are executed.
- Safety & Audit: Proposals only name node kinds from an allowlist and carry no executable code, preventing policy evasion. Everything lands on a single append-only JSONL trace, ensuring the dashboard cannot disagree with the audit trail.
MIT licensed, runs fully local via Ollama or against OpenAI/Claude APIs.
Related event: GraphARC Open Source: Deterministic Interception for Agent Workflows(4 posts)→
More from coding & agent
- Meta Launches Muse Spark 1.2 Model and Muse Code Terminal Agent — alexandr_wang · 2026-08-06
- OpenAI's Codex CLI is Open Source, Surpasses 100k GitHub Stars — dkundel · 2026-08-06
- Solving Tool Overload in AI Agents: Skill-Aware Decomposition — bendee983 · 2026-08-06
- MemoryPlugin Launches Mac Sync App, Unifying Memory for Claude Code and Other Coding Agents — testingcatalog · 2026-08-06
- MemoryPlugin Launches macOS App to Sync Local Coding Agent Chats — testingcatalog · 2026-08-06
- AI Vulnerability Scanning Reshapes Security: Bug Bounty Prices Drop as SDLC Integrates AI — philvenables · 2026-08-06