registry / @aspectlylabs/mikka / 0.0.342
@aspectlylabs/mikka@0.0.342
Local Mikka daemon: runs agents through the user's own CLIs
AI Security Review
scanned 2d ago · by lpm-firewall-aiThe inspected source routes credentials to the chosen API provider, using official defaults when no custom provider is selected, and prevents redirects. The scanner's credential-redirection and command-execution leads were not confirmed as attacks in the reviewed paths.
Static reason
One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source
Trigger
A user runs the daemon or starts an agent run using a selected provider credential or enabled mailbox.
Impact
Expected service traffic for agent runs; no unrelated credential exfiltration or concealed execution was identified.
Mechanism
The broker forwards API requests to the selected provider origin, while mailbox requests require an assigned endpoint matching the trusted run endpoint.
AI rationale
The inspected source routes credentials to the chosen API provider, using official defaults when no custom provider is selected, and prevents redirects. The scanner's credential-redirection and command-execution leads were not confirmed as attacks in the reviewed paths.
Evidence
main.jshttp-link.jsagents/native-container.mjs
Network endpoints2
api.anthropic.comapi.openai.com/v1
Decision evidence
public snapshotAI called this Clean at 91.0% confidence as Benign with low false-positive risk.
Evidence for AI clean decision
- The package is a local daemon that runs agents through the user's own CLIs.
- The API-key broker uses the official Anthropic and OpenAI API URLs as defaults.
- The broker sends a run's API key to the configured provider origin, with redirects disabled.
- The API-key broker is created with the run's selected API credential.
- The mailbox requires its assigned URL to match the trusted run endpoint before sending its run-scoped token.
Evidence against
- The scanner flags credential redirection and shell execution, but the inspected source does not substantiate either as an unrelated or concealed credential transfer.
- The changed native-container source starts the configured CLI and pipes its output; inspected code does not show unrelated outbound collection.
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
4601ai: "7.0.142",
L4602: execa: "10.1.1",
L4603: pino: "10.4.0",
High
6385assertSafeId("run id", runId);
L6386: const common = resolve4(await worktreeGitDir(repoPath, worktree), "..", "..");
L6387: if (basename(common) !== ".git") throw new Error(`${repoPath} is not a standard clone`);
...
L6396: }
L6397: async function pinnedGit(repoPath, worktree, env = process.env) {
L6398: const gitDir = await worktreeGitDir(repoPath, worktree);
L6399: const workTree = realpathSync4(worktree);
L6400: return (args, extra, options) => execa3("git", [...SAFE_GIT, ...args], {
L6401: cwd: workTree,
Findings
1 Critical5 High2 Medium4 Low
CriticalPrevious Version Dangerous Deltaagents/native-container.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@0.0.342.
See all versions of @aspectlylabs/mikka for other recorded verdicts.
- Review the evidence and your use of @aspectlylabs/mikka@0.0.342 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.