Trustra Escrow

SuspiciousAudited by ClawScan on May 10, 2026.

Overview

This escrow skill is coherent and not clearly malicious, but it gives an agent high-impact control over crypto payments and wallet credentials without clear approval limits or safeguards.

Only install or use this skill if you are comfortable giving the agent access to a Trustra-managed wallet and API key for real USDC escrow actions. Do not fund the wallet until you understand who controls signing, how disputes are resolved, and when funds can be released. Require manual approval for every payment, release, withdrawal, cancellation, dispute, and private-key export.

Findings (4)

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.

What this means

A mistaken or overly autonomous agent action could move USDC into escrow or trigger other escrow state changes without an extra human approval checkpoint.

Why it was flagged

Calling the payment script immediately sends an executing API request using the stored API key. The script does not include an in-command confirmation step, amount preview, or counterparty review before the financial action.

Skill content
payload = {"execute": True}
    
    response = requests.post(url, json=payload, headers=headers)
Recommendation

Require explicit user approval before pay, confirm, cancel, withdraw, dispute, deliver, or key-export actions. Show amount, counterparty, escrow ID, and final effect before execution, and consider disabling autonomous invocation for this skill.

What this means

Anyone or any agent with access to the API key or command output could potentially gain control of the wallet and funds.

Why it was flagged

The skill can use the Trustra API key to request and print the managed wallet private key. That makes the API key a highly privileged wallet credential, not just a service login token.

Skill content
url = f"{BASE_URL}/wallet/export"
...
print(f"Private Key: {result.get('privateKey')}")
Recommendation

Treat the Trustra API key as a wallet-control credential. Store it securely, avoid exporting private keys through agent-visible output, and require strong confirmation or separate authentication for key export.

What this means

A user may assume the escrow is fully trustless or self-custodied when the artifacts show reliance on Trustra-operated services and managed wallet credentials.

Why it was flagged

The skill uses 'trustless' framing while also relying on a managed wallet, Trustra API key, and Trustra dispute resolution. These details are disclosed, but users should understand the actual trust model before funding the wallet.

Skill content
Trustless USDC escrow for agent-to-agent transactions on Solana.
...
If problem: Either party can DISPUTE → Trustra resolves
...
Creates a managed wallet + API key stored in `credentials.json`.
Recommendation

Clarify the custody, signing, dispute-resolution, and API-key trust model prominently before users deposit SOL or USDC.

What this means

Future installs could use a different dependency version than the one originally tested.

Why it was flagged

The Python dependency is specified as a minimum version rather than pinned to an exact version. This is common, but for a financial wallet skill it reduces reproducibility of the runtime environment.

Skill content
requests>=2.28.0
Recommendation

Use a pinned dependency lockfile or exact version for financial tooling, and install only from trusted package indexes.