This release includes 1 security fix for security teams reviewing exposed deployments.
Topics
+8 more
Affected surfaces
ReleasePort's take
Moderate signalRelease v1.0.7 adds workflow automation scheduling and loop nodes while fixing skill‑script execution URL refreshes and team‑space data isolation issues.
Why it matters: Introduces schedule/loop capabilities for automating workflows; patches critical skill‑execution URL handling and resolves team‑space data sharing bugs, improving reliability.
Summary
AI summaryAdded workflow automation scheduling, loop nodes, skill script sandbox execution, and team-space ownership validation.
Changes in this release
| Type | Severity | Summary | CVE |
|---|---|---|---|
| Feature | Medium |
Support Skill script execution in a third-party sandbox from Agent nodes. Support Skill script execution in a third-party sandbox from Agent nodes. Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Feature | Medium |
Support workflow file artifacts generated by sandboxed Skill scripts. Support workflow file artifacts generated by sandboxed Skill scripts. Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Feature | Medium |
Persist Ragflow dataset ID and route by ID. Persist Ragflow dataset ID and route by ID. Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Feature | Medium |
Add workflow automation scheduling and workflow loop node capability. Add workflow automation scheduling and workflow loop node capability. Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Dependency | Medium |
Updated RPA configuration for deployment/runtime consistency. Updated RPA configuration for deployment/runtime consistency. Source: llm_adapter@2026-05-21 Confidence: low |
— |
| Dependency | Low |
Update RPA configuration for consistent deployment and runtime behavior. Update RPA configuration for consistent deployment and runtime behavior. Source: granite4.1:30b@2026-05-22-audit Confidence: low |
— |
| Performance | Medium |
Keep traceback information when sending errors, improving troubleshooting visibility. Keep traceback information when sending errors, improving troubleshooting visibility. Source: llm_adapter@2026-05-21 Confidence: low |
— |
| Performance | Low |
Retain traceback information in error messages to improve troubleshooting visibility. Retain traceback information in error messages to improve troubleshooting visibility. Source: granite4.1:30b@2026-05-22-audit Confidence: low |
— |
| Bugfix | Medium |
Fixed direct chat execution for workflow Agent nodes using Skills by refreshing Skill OSS presigned URLs before runtime execution. Fixed direct chat execution for workflow Agent nodes using Skills by refreshing Skill OSS presigned URLs before runtime execution. Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Bugfix | Medium |
Fixed workflow dialog display issues. Fixed workflow dialog display issues. Source: llm_adapter@2026-05-21 Confidence: high |
— |
| Bugfix | Medium |
Fixed team-space data sharing issues. Fixed team-space data sharing issues. Source: llm_adapter@2026-05-21 Confidence: low |
— |
| Bugfix | Medium |
Added ownership validation for space delete and update endpoints. Added ownership validation for space delete and update endpoints. Source: llm_adapter@2026-05-21 Confidence: low |
— |
| Bugfix | Medium |
Fixed CI-related issues. Fixed CI-related issues. Source: llm_adapter@2026-05-21 Confidence: low |
— |
| Bugfix | Medium |
Resolve team-space data sharing problems and add ownership validation for space update/delete APIs. Resolve team-space data sharing problems and add ownership validation for space update/delete APIs. Source: granite4.1:30b@2026-05-22-audit Confidence: low |
— |
| Refactor | Medium |
Added script sandbox provider card layout in resource management. Added script sandbox provider card layout in resource management. Source: llm_adapter@2026-05-21 Confidence: low |
— |
| Refactor | Medium |
Added clear fallback message when no script sandbox is configured. Added clear fallback message when no script sandbox is configured. Source: llm_adapter@2026-05-21 Confidence: low |
— |
Full changelog
Astron Agent v1.0.7
This release focuses on workflow reliability, automation capabilities, Skill script sandbox execution, team-space data isolation, and operational polish across the Astron Agent platform.
Highlights
- Added Skill script sandbox execution support for Agent nodes, with E2B as the first third-party sandbox provider.
- Added workflow artifact collection for Skill script outputs, allowing generated files to be uploaded to OSS and managed from the workflow file view.
- Added workflow automation scheduling support, enabling workflows to be triggered and managed by schedule.
- Added workflow loop node capability for iterative orchestration scenarios.
- Persisted Ragflow dataset IDs and routed knowledge requests by dataset ID for more stable RAG integration.
- Improved team-space data sharing behavior and strengthened ownership validation for space update/delete APIs.
New Features
- Support Skill script execution in a third-party sandbox from Agent nodes.
- Resource management now includes a script sandbox configuration entry.
- E2B is supported as the first sandbox provider through the official SDK.
- Agent nodes can expose default
run_skill_xxxtools and let the model decide when to invoke Skill scripts. - Standard
SKILL.mdformat is preserved; no runtime field is required.
- Support workflow file artifacts generated by sandboxed Skill scripts.
- Script outputs are collected from the sandbox workspace, uploaded to OSS, and associated with the workflow.
- The workflow UI includes a file entry for viewing, downloading, and deleting generated artifacts.
- Persist Ragflow dataset ID and route by ID (#1250).
- Add workflow automation scheduling and workflow loop node (#1264).
Fixes
- Fixed direct chat execution for workflow Agent nodes using Skills by refreshing Skill OSS presigned URLs before runtime execution, preventing expired
SKILL.mddownload links from causing 403 failures. - Fixed workflow dialog display issues (#1294).
- Fixed team-space data sharing issues (#1300).
- Added ownership validation for space delete and update endpoints (#1302).
- Fixed CI-related issues (#1309).
Improvements
- Added a script sandbox provider card layout in resource management, leaving room for future sandbox providers beyond E2B.
- Added a clear fallback message when no script sandbox is configured, so models can explain that administrators must configure a sandbox before Skill scripts can run.
- Keep traceback information when sending errors, improving troubleshooting visibility (#1183).
- Added governance documents (#1280).
- Updated RPA configuration for deployment/runtime consistency (#1289).
Upgrade Notes
- No breaking changes are expected for existing deployments.
- To run Skill scripts, administrators must configure a script sandbox provider in resource management. E2B is supported in this release.
- Teams using Skill scripts should verify sandbox configuration, artifact upload settings, and workflow file visibility after upgrade.
- Teams using workflow automation or loop nodes should verify workflow execution behavior after upgrade.
- Operators should pull the latest images and restart the deployment stack to apply this release.
Full Changelog
Changes are based on PRs merged into main since v1.0.6, plus the Skill sandbox execution and artifact-management work included in this release line.
Security Fixes
- Added ownership validation for space update/delete APIs, preventing unauthorized modifications
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.
Share this release
About astron-agent
Enterprise-grade, commercial-friendly agentic workflow platform for building next-generation SuperAgents.
Related context
Related tools
Beta — feedback welcome: [email protected]