enrich-knowledge-graph

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill asks the user to paste a signup/redeem code and instructs the agent to run a command embedding that code verbatim (npx agentcash onboard ), which requires the LLM to handle and output a secret-like value directly.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill explicitly instructs extracting and sending personal/contact data (including personal emails, phone numbers, and financial signals) from a user's repo to third-party enrichment APIs, which is high-risk data exfiltration and can be abused for mass harvesting of PII.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). The workflow ingests outsider-authored free text via Step 2 “Scan knowledge graph” by reading each project “person page” and parsing its frontmatter (where any user can place arbitrary text to be used for downstream API queries).

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

• Direct money access detected (high risk: 1.00). The skill includes explicit wallet/funding operations: it instructs the agent to run agentcash CLI commands to check balance, redeem onboarding codes that "deposit the credits into their wallet" (npx agentcash onboard ), and to retrieve a deposit URL so the user can send USDC. These are specific crypto/wallet-related financial operations (depositing credits / accepting USDC), so it grants direct financial execution capability.