post-engineer-mcp@1.31.0
Post Engineer MCP Server for AI pair programming and agentic workflows
AI Security Review
scanned 25d ago · by lpm-firewall-aiThe client accepts an unrestricted HTTPS endpoint override while forwarding its service API key, creating a conditional credential exposure risk. Default routing is to the package service, and telemetry is limited to tool event data with secret-like property values redacted, so the evidence supports a warning rather than a block.
Decision evidence
public snapshot- package.json defines only a prepublishOnly build hook; no install-time hook is present.
- dist/index.js reads POST_ENGINEER_API_KEY and uses it to construct the API client.
- dist/client.js accepts POST_ENGINEER_API_URL as a configurable destination and sends the API key as a bearer authorization header; non-loopback HTTP is rejected, but HTTPS destinations are not restricted.
- dist/analytics.js sends tool invocation events to the default PostHog host when POSTHOG_API_KEY is set and redacts values for secret-like property keys.
- The API defaults to https://post-engineer.com and the configurable API destination requires HTTPS except for loopback staging.
- The analytics event wrapper passes the tool name; the analytics module redacts secret-like property values before capture.
Source & flagged code
2 flaggedPackage metadata claims a different repository identity while copied source loads a runtime dependency bridge.
dist/analytics.jsView on unpkg · L61package.json defines only a prepublishOnly build hook; no install-time hook is present.
package.jsonView on unpkg · L22Findings
Affected versions and remediation
This report applies to post-engineer-mcp@1.31.0.
See all versions of post-engineer-mcp for other recorded verdicts.
- Review the evidence and your use of post-engineer-mcp@1.31.0 before allowing it. Restrict the permissions described in this report.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.