registry / @aspectlylabs/mikka / 0.0.414
@aspectlylabs/mikka@0.0.414
Local Mikka daemon: runs agents through the user's own CLIs
AI Security Review
scanned 1d ago · by lpm-firewall-aiThe flagged credential flow is a provider API broker: it uses the run credential's HTTPS base URL, restricts request paths, and rejects redirects. The new MCP file's shell capability is limited to explicitly invoked, host-configured verification checks; no malicious behavior was found.
Static reason
One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source
Trigger
A run uses a configured model API key, or an agent explicitly calls the configured verification tool.
Impact
No unauthorized credential redirection, data collection, or automatic shell execution was identified.
Mechanism
The broker forwards the key to the configured HTTPS provider origin. Verification runs only a host-configured frozen command or an inferred command when the host enables inference.
AI rationale
The flagged credential flow is a provider API broker: it uses the run credential's HTTPS base URL, restricts request paths, and rejects redirects. The new MCP file's shell capability is limited to explicitly invoked, host-configured verification checks; no malicious behavior was found.
Network endpoints2
api.anthropic.comapi.openai.com
Decision evidence
public snapshotAI called this Clean at 91.0% confidence as Benign with low false-positive risk.
Evidence for AI clean decision
- The full manifest has no install, preinstall, or postinstall lifecycle scripts.
- The API-key broker accepts an HTTPS provider URL from the run credential and checks that its protocol matches the requested provider.
- The broker forwards the API key to that configured origin, restricts requests to provider API paths, and disables redirects.
- The MCP verification tool is available only when configured for a Factory check and validates the command against host-provided settings before execution.
Evidence against
- The package provides a local daemon for running agents through the user's own CLIs, consistent with its declared purpose.
- The broker sends credentials to the configured model provider rather than a fixed publisher-controlled destination.
- Verification commands run through an explicit MCP tool call, not automatically during installation or import.
Behavioral surface
ChildProcessCryptoEnvironmentVarsFilesystemNetworkShell
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
4637ai: "7.0.142",
L4638: execa: "10.1.1",
L4639: pino: "10.4.0",
High
6566assertSafeId("run id", runId);
L6567: const common = resolve4(await worktreeGitDir(repoPath, worktree), "..", "..");
L6568: if (basename(common) !== ".git") throw new Error(`${repoPath} is not a standard clone`);
...
L6577: }
L6578: async function pinnedGit(repoPath, worktree, env = process.env) {
L6579: const gitDir = await worktreeGitDir(repoPath, worktree);
L6580: const workTree = realpathSync4(worktree);
L6581: return (args, extra, options) => execa3("git", [...SAFE_GIT, ...args], {
L6582: cwd: workTree,
Findings
1 Critical5 High2 Medium4 Low
CriticalPrevious Version Dangerous Deltamikka-mcp.mjs
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
Affected versions and remediation
This report applies to @aspectlylabs/mikka@0.0.414.
See all versions of @aspectlylabs/mikka for other recorded verdicts.
- Review the evidence and your use of @aspectlylabs/mikka@0.0.414 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.