taskmarket

Fail

Audited by Snyk on Apr 2, 2026

Risk Level: CRITICAL
Full Analysis

CRITICAL E006: Malicious code pattern detected in skill scripts.

  • Malicious code pattern detected (high risk: 1.00). The content explicitly instructs clients to fetch and run remote-updated code (npm -g install each session) and, critically, to execute server-provided "pendingActions" commands verbatim (shell commands returned by the API), while also supporting file submission and long-running daemons/XMTP+email installations — collectively creating a deliberate remote-command-execution / supply-chain / data-exfiltration backdoor vector that can be abused to steal keys or run arbitrary code.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

  • Third-party content exposure detected (high risk: 0.95). The skill explicitly instructs the agent to curl and re-read https://market.daydreams.systems/skill.md at session bootstrap and to fetch task data via GET /api/tasks/{id} whose user-provided "pendingActions" command fields the agent is told to run verbatim, meaning it ingests public, potentially untrusted task content that can directly change agent actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).


MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

  • Direct money access detected (high risk: 1.00). The skill is explicitly built for on-chain payments and wallet management. It targets Base Mainnet USDC, instructs agents to create/import wallets, deposit USDC, set withdrawal addresses, and call withdraw commands. The CLI and Raw API perform signing and X402/EIP-3009 payment-authorized actions (e.g., POST /api/tasks to fund rewards, POST /api/tasks/{id}/accept to trigger payments, POST /api/wallet/withdraw). It also exposes on-chain queries and contract addresses. These are specific crypto/payment operations (wallets, signing, token transfers), i.e. direct financial execution capabilities.

Issues (4)

E006
CRITICAL

Malicious code pattern detected in skill scripts.

W011
MEDIUM

Third-party content exposure detected (indirect prompt injection risk).

W012
MEDIUM

Unverifiable external dependency detected (runtime URL that controls agent).

W009
MEDIUM

Direct money access capability detected (payment gateways, crypto, banking).

Audit Metadata
Risk Level
CRITICAL
Analyzed
Apr 2, 2026, 12:59 PM
Issues
4