GHSA-5474-4w2j-mq4c
Claude SDK for TypeScript: Memory Tool Path Validation Allows Sandbox Escape to Sibling Directories
Packages
| Ecosystem | Package | Affected | Fixed in |
|---|---|---|---|
| npm | @anthropic-ai/sdk | >= 0.79.0, < 0.81.0 | 0.81.0 |
Details
As written in the advisory (Markdown source, shown as plain text).
The local filesystem memory tool in the Anthropic TypeScript SDK validated model-supplied paths using a string prefix check that did not append a trailing path separator. A model steered by prompt injection could supply a crafted path that resolved to a sibling directory sharing the memory root's name as a prefix, allowing reads and writes outside the sandboxed memory directory.
Users on the affected versions are advised to update to the latest version.
Claude SDK for TypeScript thanks [hackerone.com/nicksim](https://hackerone.com/nicksim) for reporting this issue!
References
External links from the advisory. They lead to third-party sites.
- https://github.com/anthropics/anthropic-sdk-typescript/security/advisories/GHSA-5474-4w2j-mq4c
- https://nvd.nist.gov/vuln/detail/CVE-2026-34451
- https://github.com/anthropics/anthropic-sdk-typescript/commit/0ac69b3438ee9c96b21a7d3c39c07b7cdb6995d9
- https://github.com/anthropics/anthropic-sdk-typescript
- https://github.com/anthropics/anthropic-sdk-typescript/releases/tag/sdk-v0.81.0