agentspan

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill documentation includes examples that embed API keys/tokens directly in shell/CLI commands and shows literal token-like strings (e.g., sk-..., ghp_..., sk_live_...), which forces an agent or output to handle or emit secrets verbatim.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The SKILL.md includes an api_tool example that sets url="https://api.stripe.com/openapi.json", which would be fetched at runtime to auto-discover an OpenAPI spec and thereby dynamically generate tools that can directly shape the agent's behavior.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill explicitly includes server-side API tooling for payment gateways (an api_tool example pointed at Stripe with STRIPE_KEY) and an example @tool that performs fund transfers (transfer_funds) which calls process_transfer and is set up for human approval. It also documents storing Stripe API keys via the credential manager. These are specific, explicit financial execution capabilities (payment API + fund transfer), not generic tooling.