acn

ReviewAudited by ClawScan on May 13, 2026.

Overview

This appears to be a legitimate ACN integration, but it handles high-value API and wallet credentials and the on-chain script can send a Base mainnet transaction by default without an in-script confirmation.

Install only if you intend to connect your agent to ACN. Before using it, set secrets through environment variables or a secrets manager, verify the ACN endpoint, avoid passing keys on the command line, use a low-value wallet, test on Base Sepolia first, and require explicit approval before any mainnet registration, payment, broadcast, or task-changing action.

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

If the agent or user runs the helper with a funded wallet, it can spend gas and permanently publish an on-chain registration.

Why it was flagged

The helper defaults to Base mainnet and directly broadcasts a signed transaction when run. On-chain registration is purpose-aligned, but mainnet blockchain writes are costly/irreversible and the code shown does not prompt for confirmation before sending.

Skill content
parser.add_argument("--chain", default="base", choices=list(CHAIN_CONFIGS), ...)
...
tx_hash_bytes = w3.eth.send_raw_transaction(signed.raw_transaction)
Recommendation

Use testnet first, verify the chain and RPC URL, and require explicit user approval before running the registration script on mainnet.

What this means

Compromise or overuse of these credentials could let someone act as the agent, change platform resources, or control the registration wallet.

Why it was flagged

The skill requires an ACN API key and can optionally use an admin-scoped Auth0 JWT and Ethereum private key. This is consistent with ACN operations, but these are high-value credentials.

Skill content
Required env: ACN_API_KEY ... Optional env: AUTH0_JWT ... WALLET_PRIVATE_KEY ... writes WALLET_PRIVATE_KEY to .env (mode 0600).
Recommendation

Use least-privilege credentials, avoid shell argv for secrets, rotate keys if exposed, and do not use a wallet holding operational funds.

What this means

Installing or running the CLI delegates trust to the npm package and its dependency chain.

Why it was flagged

The recommended CLI path runs or installs an external npm package. That is normal for this integration, but it is remote code outside the reviewed skill files.

Skill content
npx @acnlabs/acn-cli <command>
# or: npm install -g @acnlabs/acn-cli
Recommendation

Install only from the expected publisher, consider pinning versions, and review package provenance before use in sensitive environments.

What this means

Messages, task instructions, or external agent content could influence the local agent if treated as trusted instructions.

Why it was flagged

The skill supports inter-agent messaging and external A2A bridging, including public subnets. This is central to ACN, but peer messages and tasks may come from untrusted agents.

Skill content
ACN-side agents reach external agents via the gateway ... The subnet's security_schemes controls who can join — public subnet (no auth), bearer token, or API key.
Recommendation

Use manifest/allowlist modes for sensitive agents and treat received peer content as untrusted data unless the sender and task are verified.