@m8t-stack/cli@0.2.14
CLI for managing m8t-stack deployments on Azure / Microsoft Agent Foundry
AI Security Review
scanned 4mo ago · by lpm-firewall-aiStatic inspection shows a broad operator CLI with powerful but documented, user-triggered Azure/GitHub actions and no lifecycle execution, hidden exfiltration, persistence, destructive install behavior, or unconsented AI-agent control-surface mutation. Scanner findings are explained by package-aligned management features.
Decision evidence
public snapshot- dist/cli.js uses child_process spawn/spawnSync for az, gh, git, pnpm, brain-eval/brain-exam, and browser open commands.
- dist/cli.js reads Azure Key Vault GitHub App secrets and mints GitHub installation tokens.
- dist/cli.js user-invoked brain link/create writes .m8t/brain.yaml and deploys Foundry agent versions with MCP brain tooling.
- package.json has no install/postinstall/prepare hook; only prepublishOnly builds before publishing.
- dist/cli.js main only dispatches Clipanion CLI commands from argv; no import-time network or command execution beyond CLI startup registration.
- Network calls target package-aligned Azure, GitHub, GHCR, OpenAI SDK, Copilot MCP, and configured gateway endpoints for deployment/management workflows.
- Local file writes are expected CLI state/output such as ~/.m8t-stack configs, GitHub App creds, temp brain repos, eval outputs, and project env files.
- README.md documents the Azure/GitHub management purpose and local config paths matching source behavior.
Source & flagged code
2 flagged · loading sourceSource combines command execution, command-output handling, and outbound requests; review data flow before blocking.
dist/cli.jsView on unpkg · L19448Findings
Affected versions and remediation
This report applies to @m8t-stack/cli@0.2.14.
See all versions of @m8t-stack/cli for other recorded verdicts.
- Review the evidence and your use of @m8t-stack/cli@0.2.14 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.