The AI review rates the findings as likely false positive, but the risk score (100/100) still counts them.
Analysis record
- Analysed
- 6 months ago 1 May 2026
- Version
- v1.2.22
- Artifact
- SHA256 850088f1…e75e3b
- Detail
- Partial list
Evidence ledger
Ranked by severity · findings with a source location link to the code viewer
Finding Categories
MCP Server Analysis
MCP servers expose tools and resources to AI assistants. Unlike browser extensions, they run as standalone processes with direct system access. Tool definitions are analyzed for prompt injection, data exfiltration, and tool poisoning patterns.
AI Security Report
AI Security Review
Evidence context: threat category none; evidence quality moderate.
Tool Poisoning Assessment
The evidence shows only 1 tool-poisoning finding in a package with 4,146 total findings. This is consistent with legitimate MCP tool definition code triggering the scanner. The single finding lacks evidence of hidden AI manipulation directives (e.g., "do not tell the user", "silently execute", XML-style instruction tags). Without seeing the actual finding content, the pattern indicates a false positive from tool registration code rather than malicious tool poisoning.
Credential and Network Access
The 39 secret findings are code-smell detections, not credential theft. No findings target sensitive paths like .ssh/, .aws/credentials, .kube/config, or specific secret names (GITHUB_TOKEN, AWS_SECRET_ACCESS_KEY). The 20 network findings are all in expected locations:
dist/extension.js:56573anddist/cli.js:84917are bundled distribution files containing standard fetch callssrc/webview/standalone-shim.ts:108andsrc/webview/standalone-shim.ts:243are webview shim code for the standalone web appsrc/standalone/__tests__/server.integration.test.ts:89is a test file
No network findings show connections to suspicious external domains. The combination of credential reads + unknown domain exfiltration (the defining exfiltration pattern) is absent.
IoC Extractor Noise
All 3,088 IoC findings are false positives from the XIOC extractor misreading code as domains:
currentdocument.saveis a property access chain, not a domainrenderroutenotfound.js.mapandrouterprovider.js.mapare source map filenamescard-markdown-tab.is,comment-editor-tab.is,p.is,cm-editor.cmare file extensions (.is for Island components, .cm for CodeMirror)
This is documented scanner noise. The 672 malware-signature findings with 0 actual malware detections confirms bundled dependencies are generating noise.
Strongest Counterargument
The high finding count (4,146) and 672 malware signatures could suggest malicious code. However, this pattern is expected for MCP packages with bundled dist/ files and standard dependencies. The 0 malware findings, 0 credential-access findings targeting sensitive paths, and 0 suspicious domain connections confirm these are false positives from bundled code, not evidence of malicious behavior.
Conclusion
This is a legitimate Kanban board management tool (VSCode extension, CLI, MCP server, standalone web app) with no evidence of tool poisoning, credential exfiltration, or other malicious behavior. All findings are explainable as bundled dependency noise and scanner false positives.
Key Reasons
- All 3088 IoC findings are false positives (property chains, file extensions, source maps misread as domains)
- Network findings are in dist/ bundled files and test files with no suspicious domains
- No credential-access findings targeting sensitive paths (.ssh, .aws, .kube)
- 672 malware signatures with 0 actual malware detections indicates bundled dependency noise
- Single tool-poisoning finding lacks evidence of hidden AI manipulation directives
False Positive Considerations
- XIOC extractor misreading property chains and file extensions as domains
- Bundled dist/ files generating multiplicative false positives
- Malware signature rules matching benign bundled dependencies
- Code-smell findings from standard Node.js patterns
Reviewed 2026-04-27; recommended action: suppress false positive; model confidence 85%.
MCP version history
Risk trend by version
2 analyzed versions. Each point is the latest successful scan for that version; failed zero-score scans are hidden. Dates: First seen by Risky Plugins.
Pick any point on the chart, then open the Source tab to read that version's code.
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