ckb-ccc-fundamentals

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

• Third-party content exposure detected (high risk: 0.75). SKILL.md instructs the runtime workflow to fetch documentation pages/indices from external URLs (e.g., https://docs.ckbccc.com/llms.txt, per-page .md, and llms-full.txt), which are outsider-authored free text that can be ingested into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill instructs the agent at runtime to fetch authoritative docs (e.g., GET https://docs.ckbccc.com/llms.txt) and per-page Markdown (https://docs.ckbccc.com/en/docs//.md) which are used to ground generated code and signatures, so these external URLs directly control the agent's prompts/context.

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

• Direct money access detected (high risk: 1.00). The skill is a hub for the CKB blockchain SDK and explicitly references signer/wallet setup and "composing/sending a transaction" (and includes signer APIs such as signer.getRecommendedAddress and amount-conversion helpers). These are concrete crypto/blockchain capabilities (wallets/signers and transaction-sending) that enable on-chain transfers — i.e., direct financial execution.