Granite Code Assistant
The AI review rates the findings as likely false positive. The risk score (76/100) comes from stored counts that the review could not adjust. A new analysis will update it.
Analysis record
- Analysed
- 11 months ago 9 December 2025
- Version
- v0.1.2024102233
- Artifact
- SHA256 d6671698…5ac19f
- Detail
- Counts only, rows expired
Detail for this analysis has expired
This analysis ran on 9 December 2025. 35 findings counted, and the counts are kept. The individual rows have aged out of storage. A new analysis restores them.
Detail for this analysis has expired
This analysis ran on 9 December 2025. 536 network indicators counted, and the counts are kept. The individual rows have aged out of storage. A new analysis restores them.
Publisher Evidence
LowRed Hat
Publisher identity, store signals, distribution reach, and warning signals used for context. Treat this as supporting evidence, not a clean bill of health.
11 evidence rows available.
Detail for this analysis has expired
This analysis ran on 9 December 2025. 571 findings counted, and the counts are kept. The individual rows have aged out of storage. A new analysis restores them.
Stored counts by category
- 536
- Network indicators
- 31
- Code patterns
- 31
- Malware signatures
- 2
- Dependencies
- 2
- Network calls
AI Security Report
AI Security Review
Evidence context: threat category none; evidence quality moderate.
The Granite Code Assistant extension from Red Hat requires filesystem read access to analyze source code and provide AI-powered code suggestions. This access is justified by the extension's stated purpose: "Use IBM Granite Code LLM as your Code Assistant." The extension reads workspace files to send code context to the Granite LLM backend, which is the core functionality of any code assistant.
The 536 IoC findings in this bundle are XIOC extractor false positives. Specific findings like XIOC-DOMAIN-b.call, XIOC-DOMAIN-this.cbs, XIOC-DOMAIN-fe.lt, and XIOC-DOMAIN-fe.tab are JavaScript property access chains being misread as domain names. The pattern b.call matches the documented false-positive pattern where property access chains like b.call, h.next, g.id are incorrectly flagged as domains. These are not actual network destinations but minified JavaScript variable access patterns.
The 2 network findings are legitimate webview communication. NET-SOCKET_IO-extension/webviews/build/assets/index.js-49 and NET-FETCH-extension/webviews/build/assets/index.js-1 are in the webview build output directory. A code assistant extension must communicate with its LLM backend via network calls, and socket.io is a standard real-time communication library. These network calls occur in extension/webviews/build/assets/index.js, which is the compiled webview bundle where user interface code communicates with the extension host.
There are zero secret or credential findings. The findings summary shows "secret":"0" in the threat_indicators. This means the extension does not access .env files, SSH keys, cloud credentials, or VS Code secret storage. The 31 code-smell findings are low-severity code quality patterns that match basic Node.js patterns and are classified as noise per the known false-positive patterns.
The strongest counterargument is the low user count of 22, which might suggest a newly published or suspicious extension. However, Red Hat is a verified publisher on the VS Code marketplace, and the version number 0.1.2024102233 indicates this is an early release from October 2024. New extensions from verified publishers have lower user counts during initial rollout. The combination of verified publisher status, zero credential findings, and documented false-positive patterns in all IoC findings outweighs the low user count concern.
This extension exhibits expected behavior for an AI code assistant: reading workspace files for code analysis, spawning webview processes for the UI, and making network calls to the LLM backend. The 571 total findings are inflated by XIOC extractor garbage (536 findings) and code-smell noise (31 findings), leaving only 2 legitimate network findings that are justified by the extension's purpose.
Key Reasons
- All 536 IoC findings are XIOC extractor false positives from property access chains
- Verified publisher (Red Hat) on VS Code marketplace
- Zero credential or secret findings
- Network activity in webview build assets is expected for AI code assistant
- No malware signatures or obfuscation findings
False Positive Considerations
- XIOC property access chains misread as domains (b.call, this.cbs, fe.lt)
- Bundled webview build output triggering network detection
- Code-smell findings on legitimate Node.js patterns
Reviewed 2026-04-29 · version 0.1.2024102233; recommended action: suppress false positive; model confidence 85%.
Source Code Not Available
Source code is not available for this version of the extension.
About This Extension
Frequently Asked Questions
Similar Extensions
Related extensions from the same publisher or marketplace
Migration Toolkit for Applications (MTA) by Red Hat
Red-Hat
Red Hat OpenShift Dev Spaces
Red-Hat
Red Hat Dependency Analytics
Red-Hat
Quarkus Tools
Red-Hat
OpenShift Dev Spaces
Red-Hat
Server Connector by Red Hat
Red-Hat