registry / @aspectlylabs/mikka / 0.0.367
@aspectlylabs/mikka@0.0.367
Local Mikka daemon: runs agents through the user's own CLIs
AI Security Review
scanned 2d ago · by lpm-firewall-aiThe suspicious routing hints are consistent with normal agent execution and credential brokering. The inspected code sends model credentials to the official provider by default or to a user-configured HTTPS provider, and the native runner routes requests through a local broker.
Static reason
One or more suspicious static signals were detected.
Trigger
Users invoke the daemon or start an agent run; no install-time hook is declared.
Impact
No unauthorized credential transfer or other malicious effect was identified in the inspected flows.
Mechanism
Agent model requests pass through a local broker, which sends them to the official provider by default or to a configured HTTPS provider.
AI rationale
The suspicious routing hints are consistent with normal agent execution and credential brokering. The inspected code sends model credentials to the official provider by default or to a user-configured HTTPS provider, and the native runner routes requests through a local broker.
Evidence
main.jsagents/native-container.mjsgithub-credentials.mjs
Network endpoints2
api.anthropic.comapi.openai.com
Decision evidence
public snapshotAI called this Clean at 94.0% confidence as Benign with low false-positive risk.
Evidence for AI clean decision
- The manifest describes a local daemon that runs agents through the user's own CLIs and declares no lifecycle scripts.
- Model credentials use the official Anthropic or OpenAI API by default; custom provider URLs must use HTTPS.
- The API-key broker attaches the credential to requests sent to the configured provider origin for model API paths.
- The native runner routes model traffic through a local proxy to the mounted task broker socket.
- GitHub run credentials are obtained from a configured endpoint using a run token and cached with restrictive file permissions.
Evidence against
- The package executes agent CLIs and performs network requests as part of its stated purpose.
- The broker handles model API credentials, but the inspected flow sends them to the configured model provider.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetworkShell
HighEntropyStrings
Source & flagged code
6 flaggedgithub-credentials.mjsView file
7import { randomUUID } from 'node:crypto'
L8: import { spawn } from 'node:child_process'
L9: import { pathToFileURL } from 'node:url'
High
Child Process
Package source references child process execution.
github-credentials.mjsView on unpkg · L7mikka.jsView file
4625ai: "7.0.142",
L4626: execa: "10.1.1",
L4627: pino: "10.4.0",
High
6446assertSafeId("run id", runId);
L6447: const common = resolve4(await worktreeGitDir(repoPath, worktree), "..", "..");
L6448: if (basename(common) !== ".git") throw new Error(`${repoPath} is not a standard clone`);
...
L6457: }
L6458: async function pinnedGit(repoPath, worktree, env = process.env) {
L6459: const gitDir = await worktreeGitDir(repoPath, worktree);
L6460: const workTree = realpathSync4(worktree);
L6461: return (args, extra, options) => execa3("git", [...SAFE_GIT, ...args], {
L6462: cwd: workTree,
Findings
5 High2 Medium5 Low
HighChild Processgithub-credentials.mjs
HighShellmikka.js
HighSame File Env Network Executionmikka.js
HighCredential Redirect Persistencemain.js
HighCommand Output Exfiltrationagents/native-container.mjs
MediumNetwork
MediumEnvironment Vars
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
LowCopyleft License
LowSuspicious Lifecycle Evidencepackage.json
Affected versions and remediation
This report applies to @aspectlylabs/mikka@0.0.367.
See all versions of @aspectlylabs/mikka for other recorded verdicts.
- Review the evidence and your use of @aspectlylabs/mikka@0.0.367 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.