0xwork

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

• Third-party content exposure detected (high risk: 1.00). The skill's required workflow (SKILL.md) tells the agent to fetch and read user-generated task descriptions, comments, and referenced URLs from the public 0xWork marketplace (via commands like "0xwork discover" and "0xwork task " and tools like web_search/web_fetch in references/execution-guide.md), which are explicitly untrusted and can contain instructions that could influence agent actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs running the CLI at runtime (e.g., "npx @0xwork/cli discover") which fetches and executes code from the npm package URL https://npmjs.com/package/@0xwork/cli, so remote code would be executed as a required dependency.

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

• Direct money access detected (high risk: 1.00). The skill explicitly provides blockchain payment and wallet capabilities. It includes wallet configuration (PRIVATE_KEY or BANKR_API_KEY for remote signing), env vars for RPC and API, and CLI/SDK commands that perform on-chain financial actions: posting tasks with USDC bounties (deposits to escrow), claiming tasks (staking $AXOBOTL), approving/releasing USDC, product buy (purchasing with USDC), faucet/funding, reclaiming bounties, and other transaction-affecting commands. It also references smart contract addresses and a Bankr remote-signing API. These are specific crypto/blockchain payment and signing functions (not generic HTTP or browser tools), so it grants direct financial execution authority.