Disclosed weakness · no CVE recorded
Workspace skill import could expose a server directory to an AgentScope agent
AgentScope · AgentScope · workspace skills
First documented disclosure · Review checkpoint
Summary
An HTTP caller could import an eligible server-side directory into an agent workspace and read its contents there.
Severity and exploitation evidence
No CVSS score recorded in this review.
Exploitation not established: The public report and upstream fix document the defect. They identify no malicious use.
Technical severity is not a measure of observed attacks or of risk to your deployment. Attempts, demonstrations, and successful compromise retain the qualifications in the source.
How the weakness works
The importer confined the destination while accepting a filesystem source outside approved skill directories.
Conditions required
A caller can reach the affected endpoint and select a server-readable directory containing a valid skill manifest.
Potential impact
Server files accompanying that manifest could become readable within the agent workspace.
Affected versions and configurations
The project verified the pre-fix implementation. A complete affected release range was not published.
Fix and mitigation status
2.0.12 includes the correction.
Practical next steps
Upgrade to 2.0.12 or a later fixed version. Use the supported skill upload or library import and retire clients using server-path imports.
Operating lesson
Authorize both the source and destination of an import.
Evidence and limits
A July 13 public issue describes the finding and an earlier private report. October 28 PR 3536 removed the HTTP endpoint. The November 16 release lists the patch.
Sources
- AgentScope original issue
Primary source ·
- AgentScope merged correction
Primary source ·
- AgentScope 2.0.12 release notes
Primary source ·