@aspectlylabs/mikka@0.0.274
Local Mikka daemon: runs agents through the user's own CLIs
AI Security Review
scanned 2d ago · by lpm-firewall-aiThe source reviewed does not show the flagged credential-redirection or execution behavior as an attack: the API key broker restricts requests to model API paths and sends them to the selected provider with redirects disabled. The package documents the Mikka service and daemon workflow, and its source says run credentials are kept in memory and are not journaled.
Decision evidence
public snapshot- The manifest exposes the daemon CLI and library entry points; the README describes signing in and starting the Mikka daemon.
- Run API credentials are kept in memory and handed only to that run's agent process or container; the source says they are never journaled.
- The API-key broker accepts only specific Anthropic or OpenAI model request paths and forwards requests to the configured provider origin with redirects disabled.
- The CLI’s documented default service is app.aspectlylabs.com; a user can also supply a preview or self-hosted server URL.
- The broker’s default model destinations are api.anthropic.com and api.openai.com; credentials are sent with model requests to those destinations or the configured provider.
Source & flagged code
6 flaggedPackage source references child process execution.
github-credentials.mjsView on unpkg · L4Findings
Affected versions and remediation
This report applies to @aspectlylabs/mikka@0.0.274.
See all versions of @aspectlylabs/mikka for other recorded verdicts.
- Review the evidence and your use of @aspectlylabs/mikka@0.0.274 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.