Security audit
Security checks for vulnerabilities and agentic risk
The skill is a coherent BotHire marketplace pointer, but it asks agents to use unpinned remote packages and local wallet keys for autonomous payments without strong user-control boundaries.
Review before installing. Use BotHire only with explicit user approval for each paid action, avoid long-lived or high-value wallet keys, prefer an isolated hot wallet with limited funds, and verify exact package versions or source provenance before running the npx commands.
The skill instructs users to run npx clawhub@latest install bothireagent/bothire / npx clawhub install ... without pinning an exact trusted version. Unpinned package execution allows supply-chain compromise or unexpected behavior changes if the upstream package is maliciously updated, hijacked, or serves different code over time.
The invocation guidance is very broad and overlaps with common agent tasks such as research, code review, translation, data, scraping, and payments. In an automated agent ecosystem, overly generic trigger language can cause the skill to activate in many unrelated contexts, increasing the chance of unexpected external calls, payment flows, or delegation to third-party services.
The skill tells users to run npx bothire-mcp while also configuring BOTHIRE_PRIVATE_KEY and BOTHIRE_API_KEY. Because the package version is not pinned, a compromised or changed package could directly exfiltrate signing keys or misuse wallet authority, making this more dangerous than a generic unpinned install.
The instruction npx bothire executes a remote package without version pinning in a workflow that generates a wallet and registers an identity. If the package is replaced or maliciously updated, it could generate insecure wallets, capture secrets, or register users through attacker-controlled flows.
