@botbuddy/cli@1.64.2
BotBuddy — Swarm coordination CLI for multi-agent workflows
AI Security Review
scanned 4d ago · by lpm-firewall-aiThe manifest has no install hooks, and the reviewed launcher, credential transport, update check and staging-pool setup have purpose-related triggers and destinations. The scanner flags are consistent with the CLI's process, network, credential and explicit persistence capabilities; I found no concrete malicious behavior.
Decision evidence
public snapshot- The manifest exposes the expected BotBuddy CLI binaries and declares no lifecycle scripts.
- The launcher starts delivery only for ordinary CLI invocations; help and inspection paths skip it, and it does not run during package installation.
- The staging session pool is described as human-installed; its install handler checks for a human session before network or file operations.
- Wait requests send the supplied token to the configured service URL, consistent with the CLI's relay workflow.
- The update check fetches the latest package metadata from the official npm registry.
- The CLI has a delivery process and runtime telemetry-related code that merit review, but the inspected launcher and outbox material did not establish unrelated collection or credential theft.
- The package can install a launchd job and handle staging sessions, but the inspected setup is an explicit human-only command tied to the package's purpose.
Source & flagged code
7 flaggedPackage source references child process execution.
src/update-check.mjsView on unpkg · L16Source writes installer persistence such as shell profile or service configuration.
Findings
Affected versions and remediation
This report applies to @botbuddy/cli@1.64.2.
See all versions of @botbuddy/cli for other recorded verdicts.
- Review the evidence and your use of @botbuddy/cli@1.64.2 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.