road-nearest

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill explicitly instructs users to export a private key as FITTINGS_PRIVATE_KEY (e.g., export FITTINGS_PRIVATE_KEY=0x...) and notes the key sits in plaintext, which encourages placing raw secrets into shell commands/env and would require the LLM to include secret values verbatim if used that way.

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 1.00). These URLs point to a personal domain that hosts and explicitly instructs users to download/run an executable script (call.mjs), and fetching/running executables from an unverified site is a high-risk malware vector.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill instructs fetching and then running remote code (call.mjs) from https://fittings.sh/skill/call.mjs (curl -sO https://fittings.sh/skill/call.mjs), so that URL supplies executable code the agent may run.

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

• Direct money access detected (high risk: 1.00). The skill explicitly describes a metered paid API where the client must sign a payment authorization with a private key (export FITTINGS_PRIVATE_KEY), references USDC on Base, off-chain signing and facilitator settlement, and provides client tools (call.mjs, @x402/evm) to perform the payment flow. This is a specific crypto payment/wallet signing capability that can cause real spends, so it grants direct financial execution authority.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.70). The skill does not ask for sudo, system-file edits, or user creation, but it directs the agent to download and execute remote code and to store a private key in plaintext environment variables on the host—actions that meaningfully risk compromising the machine's security.