@saltpub/xiaojiu-agent-bridge
Bridge Xiaojiu IM with local CLI coding agents
This report shows the evidence for @saltpub/xiaojiu-agent-bridge@0.1.6. 1 of 6 known versions is blocked or flagged. The newest version, 0.1.7, is flagged. See all versions.
AI Security Review
scanned 2d ago · by lpm-firewall-aiThe CLI deliberately bypasses its prompt redaction and sends up to 8,192 characters of agent prompt content through default-on telemetry. This is concealed collection of user-provided content, so the package should be blocked.
Decision evidence
public snapshot- The CLI loads the package’s telemetry adapter by default; disabling it requires a specific environment variable.
- The bridge logs an agent prompt capped at 8,192 characters and deliberately names it `agentPrompt` to bypass the sanitizer that redacts `prompt`.
- Sanitized log fields are passed to the telemetry adapter’s `emit` method.
- The telemetry adapter receives the app ID and machine hostname when the bridge starts.
- The prompt is used by the bridge’s core feature: sending Xiaojiu messages to a local coding agent.
- Telemetry can be disabled with `XIAOJIU_BRIDGE_TELEMETRY_DISABLED`.
- The telemetry adapter’s network destination is not present in the inspected package source.
Source & flagged code
5 flaggedThis package version adds a dangerous source file absent from the previous stored version; route for source-aware review.
dist/cli.jsView on unpkgFindings
Affected versions and remediation
This report applies to @saltpub/xiaojiu-agent-bridge@0.1.6.
The versions list below shows the verdict for each recorded version.
- Avoid installing @saltpub/xiaojiu-agent-bridge@0.1.6. Remove it from direct dependencies and check your lockfile for transitive copies.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
- If this version ran, investigate the affected machine and build environment. Rotate credentials it could access and rebuild from a trusted environment.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.