aethel-autapon@1.0.2
Compilador DAG para coordinación de agentes autónomos en Monad
AI Security Review
scanned 6d ago · by lpm-firewall-aiThe package's main entry automatically uses locally configured private keys to submit repeated payments to a hard-coded recipient without per-payment confirmation. Automatic spending of wallet funds is malicious under the review rules.
Static reason
One or more suspicious static signals were detected.
Trigger
Importing or running dag-compiler.mjs.
Impact
Agent wallets may spend funds on payments the user did not individually request.
Mechanism
It loads private keys from mandatos-agentes.json, constructs three 0.05 MON payment intents per agent, and submits payments through the Mandate SDK.
Attack narrative
The package main entry runs a payment workflow at top level. It loads local agent private keys, creates three 0.05 MON intents per agent for a fixed recipient, and submits payments without per-payment confirmation, causing wallet funds to be spent when the entry runs.
AI rationale
The package's main entry automatically uses locally configured private keys to submit repeated payments to a hard-coded recipient without per-payment confirmation. Automatic spending of wallet funds is malicious under the review rules.
Evidence
package.jsondag-compiler.mjsmandatos-agentes.json
Network endpoints1
rpc.monad.xyz
Decision evidence
public snapshotAI called this Malicious at 98.0% confidence as Malware with low false-positive risk.
Evidence for block
- package.json sets dag-compiler.mjs as the package main entry.
- The entry reads mandatos-agentes.json and maps each agent's private key into runtime state.
- The entry creates three 0.05 MON payment intents per agent for a fixed recipient.
- The top-level workflow validates intents and then executes the payment layers without an explicit per-payment confirmation.
- Payments use the loaded agent keys through the SDK's pay method.
Evidence against
- The manifest has no install or lifecycle scripts.
- The package describes itself as a Monad agent coordination tool, where on-chain payments can be part of its stated purpose.
Behavioral surface
EnvironmentVarsFilesystemNetwork
HighEntropyStringsUrlStrings
GitDependency
Source & flagged code
5 flaggedpackage.jsonView file
2"name": "aethel-autapon",
"version": "1.0.2",
"description": "Compilador DAG para coordinación de agentes autónomos en Monad",
"main": "dag-compiler.mjs",
"type": "module",
"keywords": [
"monad",
"dag",
"
High
Ai Review Evidence
package.json sets dag-compiler.mjs as the package main entry.
package.jsonView on unpkg · L2dag-compiler.mjsView file
52const mandatos = JSON.parse(readFileSync("mandatos-agentes.json", "utf8"));
const agentes = mandatos.map(m => ({
agentId: m.agentId,
address: m.address,
mandateId: m.mandateId,
privateK
High
Ai Review Evidence
The entry reads mandatos-agentes.json and maps each agent's private key into runtime state.
dag-compiler.mjsView on unpkg · L52361const PAGOS_POR_AGENTE = 3;
const RECIPIENT = "[redacted]";
const intenciones = [];
for (const agente of agentes) {
for (let i = 0; i < PAGOS_POR_AGENTE; i++) {
intenciones.push({
agente,
amount: "0.05",
recipient: RECIPIENT,
});
}
}
console.log("Total de intencione
Findings
5 High3 Medium3 Low
HighAi Review Evidencepackage.json
HighAi Review Evidencedag-compiler.mjs
HighAi Review Evidencedag-compiler.mjs
HighAi Review Evidencedag-compiler.mjs
HighAi Review Evidencedag-compiler.mjs
MediumNetwork
MediumEnvironment Vars
MediumGit Dependency
LowFilesystem
LowHigh Entropy Strings
LowUrl Strings
Affected versions and remediation
This report applies to aethel-autapon@1.0.2.
See all versions of aethel-autapon for other recorded verdicts.
- Avoid installing aethel-autapon@1.0.2. Remove it from direct dependencies and check your lockfile for transitive copies.
- Choose an independently verified alternative or release. This report does not establish that other versions are safe.
- If this version ran, investigate the affected machine and build environment. Rotate credentials it could access and rebuild from a trusted environment.
Evidence last updated: .
Think this verdict is wrong? Dispute it on GitHub or read how reports are produced.