@mseep/mcp-server-postgres-init
From the RiskyPlugins AI security review of the observed evidence.
Analysis record
- Analysed
- 2 months ago 26 August 2026
- Version
- v1.0.0
- Artifact
- SHA256 3d2fa5cc…92cad8
- Detail
- Full list
Evidence ledger
Ranked by severity · findings with a source location link to the code viewer
YARA Rule Matches
3 rules| Severity | Rule | Hits | Files | Metadata |
|---|---|---|---|---|
| LOW | postinstall file manipulation | 4 | README.mdsrc/index.tsbuild/index.js +1 more | - |
| LOW | postinstall network communication | 3 | README.mdsrc/index.tsbuild/index.js | - |
| LOW | postinstall system command | 3 | README.mdsrc/index.tsbuild/index.js | - |
Finding Categories
YARA Rules Matched
3 rules(10 hits)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 weak.
Security Analysis: @mseep/mcp-server-postgres-init
Package Overview
This MCP server (@mseep/mcp-server-postgres-init, v1.0.0) is designed to initialize PostgreSQL databases in Docker containers. The developer is listed as skydeckai. The package has 0 users, indicating it is either newly published or has limited adoption.
Tool Poisoning Assessment
No tool-poisoning findings were detected. The findings_by_category bundle is completely empty, meaning the security scanner did not identify any hidden AI manipulation directives, XML-style instruction tags, or invisible Unicode characters in tool descriptions. This is the most critical finding for MCP servers, and the absence of such findings is positive.
Credential and Network Access
No credential-access or network-related findings were detected. The empty findings bundle indicates:
- No reads of sensitive paths (
.ssh/,.aws/credentials,.kube/config,application_default_credentials.json) - No HTTP calls to unexpected or undocumented domains
- No evidence of exfiltration architecture (credential harvesting combined with network calls to unknown endpoints)
For a PostgreSQL initialization tool, this is the expected behavior. The package should only need Docker API access and PostgreSQL configuration parameters, neither of which would trigger credential-access findings.
Evidence Quality and Limitations
The evidence_quality is weak because the findings_by_category object is completely empty. This could indicate:
- The package is genuinely minimal with no security-relevant code patterns
- The analysis did not fully execute or scan all files
- The package structure did not trigger any YARA rules or IoC extraction
Given the straightforward purpose (PostgreSQL Docker initialization), interpretation #1 is plausible. However, the absence of any findings—including typical code-smell hits that appear in almost any JavaScript package—suggests the analysis may have been incomplete.
Strongest Counterargument
The strongest argument against this verdict is the complete absence of findings. Most MCP packages, even benign ones, generate some code-smell findings from YARA rules matching common patterns (process.env, fetch, fs operations). Zero findings across all categories is unusual and could indicate incomplete scanning rather than a genuinely clean package.
This counterargument does not change the conclusion because:
- The package description is legitimate and specific
- No suspicious patterns were detected
- A simple initialization tool may genuinely have minimal attack surface
- Without evidence of malicious behavior, the default position should be benign
Recommendation
no_action is appropriate. The package shows no signs of tool poisoning, credential theft, or data exfiltration. Users should still verify the package's behavior in a controlled environment before production use, as with any MCP server.
Conclusion
This MCP server presents no detectable security risks based on the available evidence. The empty findings bundle, while unusual, combined with the legitimate package purpose and absence of any suspicious patterns, supports a benign classification.
Key Reasons
- Zero security findings across all categories
- Legitimate package purpose (PostgreSQL Docker initialization)
- No tool-poisoning or credential-access detections
- No network calls to suspicious domains
Reviewed 2026-04-24 · version 1.0.0; recommended action: no action; model confidence 75%.
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