registry / @arispay/skill / 2.1.0

@arispay/skill@2.1.0

ArisPay skill for Hermes-style AI agents. Directs agents to `npx payagent pay <url>` for any paid API call — self-bootstraps auth, agent provisioning, and USDC funding.

AI Security Review

scanned 18h ago · by lpm-firewall-ai

LPM treats this as warn-only first-party agent extension lifecycle risk. An explicit installer places this package-owned agent skill in Claude/Hermes skill directories. The installed skill directs agents to invoke a remotely resolved payment CLI; no unconsented install-time execution is present.

Static reason

One or more suspicious static signals were detected.

Trigger

User runs arispay-skill install, then an agent follows the installed payment skill.

Impact

Can influence an agent to bootstrap and use a payment capability; npx resolves and executes payagent at runtime.

Mechanism

First-party agent-skill installation plus npx payment-CLI delegation.

Rationale

This is an explicit first-party agent-extension setup with runtime delegation to a payment CLI, so it warrants warning as a dangerous capability. It does not meet the blocking threshold because installation and execution are user/agent invoked and no concrete malicious behavior was found.

Evidence

package.jsonsrc/install.mjsskill/SKILL.mdskill/scripts/pay.mjs<cwd>/.claude/skills/arispay/~/.claude/skills/arispay/~/.hermes/skills/arispay/

Decision evidence

public snapshot

AI called this Suspicious at 91.0% confidence as Dangerous Capability with low false-positive risk.

@arispay/skill@2.1.0

AI Security Review

Decision evidence

Source & flagged code

Source & flagged code

Findings

Source & flagged code