AgentHire

AdvisoryAudited by Static analysis on May 10, 2026.

Overview

Detected: suspicious.dangerous_exec

Findings (1)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

ConcernHigh Confidence

ASI02: Tool Misuse and Exploitation

What this means

The agent could spend the funded wallet’s testnet ETH and release escrow for a provider response without the user reviewing the exact cost or result first.

Why it was flagged

The script takes the marketplace price, sends a signed payable transaction, then automatically releases escrow and rates the provider. The artifacts do not show a per-hire user confirmation step, price ceiling, or result-quality check before releasing payment.

Skill content

const price = s[5]; // pricePerJob in wei ... const tx = await escrow.createJob(serviceId, task, { value: price, nonce }); ... const confirmTx = await escrow.confirmComplete(jobId, { nonce: cNonce }); ... const rateTx = await escrow.rateJob(jobId, 5, { nonce: rNonce });

Recommendation

Require explicit user approval before each hire and before escrow release, and add a maximum price/chain check in the script.

NoteHigh Confidence

ASI03: Identity and Privilege Abuse

What this means

Anyone who gets the .env private key can control the agent wallet and spend its funds.

Why it was flagged

The setup flow creates a wallet private key and persists it in .env so the skill can sign blockchain transactions. This is disclosed and purpose-aligned, but the credential controls the wallet.

Skill content

AGENTHIRE_PRIVATE_KEY=${wallet.privateKey} ... fs.writeFileSync(envPath, envContent);

Recommendation

Use only a dedicated Base Sepolia wallet with minimal funds, never reuse a mainnet or personal wallet key, and protect the .env file.

ConcernHigh Confidence

ASI07: Insecure Inter-Agent Communication

What this means

Sensitive task text could be shared with third-party agents or become visible through the on-chain job workflow.

Why it was flagged

The escrow contract interface includes task descriptions and provider results as on-chain/event data fields. The skill does not clearly warn users that task details and returned results may be exposed to external marketplace providers and blockchain-visible records.

Skill content

"name": "JobCreated" ... "name": "taskDescription", "type": "string" ... "name": "ResultSubmitted" ... "name": "result"

Recommendation

Do not include secrets or private data in tasks, and add clear warnings plus data-minimization rules before sending task descriptions to providers.

NoteHigh Confidence

ASI05: Unexpected Code Execution

What this means

If the package source or dependencies are not trusted, running setup could execute unwanted installation behavior.

Why it was flagged

Running setup.js executes a shell command to install Node dependencies. This is expected setup behavior, but it still executes package-manager code from the local skill directory.

Skill content

execSync("npm install --production", { cwd: dir, stdio: "inherit" });

Recommendation

Review the skill source and package dependencies before running setup or npm install.

Findings (1)

critical

suspicious.dangerous_exec

Location: setup.js:16
Finding: Shell command execution detected (child_process).