The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill manages agent private keys stored in ~/.openclaw/billions/kms.json. According to scripts/shared/storage/keys.js and scripts/shared/storage/crypto.js, the skill defaults to storing these keys in plaintext if the BILLIONS_NETWORK_MASTER_KMS_KEY environment variable is not provided. While the storage location is outside the workspace and permissions hardening is recommended in documentation, the default posture presents a risk of credential exposure if the host or the specific directory is compromised.
[DATA_EXFILTRATION]: The skill performs network operations to external domains for identity resolution and payment processing.
In scripts/verifySignature.js, it fetches DID documents from https://resolver.privado.id/1.0/identifiers.
In scripts/shared/utils/auth.js, it communicates with https://identity-dashboard.billions.network for URL shortening.
In scripts/buildX402Payment.js, it sends authentication tokens and payment signatures to URLs provided in the paymentRequired challenge argument. This is functional behavior for the x402 protocol but involves sending cryptographic attestations to external endpoints.
[COMMAND_EXECUTION]: The skill provides several Node.js scripts (createNewEthereumIdentity.js, signChallenge.js, etc.) that perform cryptographic operations and local file system writes to manage identity state. These scripts are intended to be executed by the agent or operator.
[EXTERNAL_DOWNLOADS]: The skill depends on standard identity and cryptographic libraries from the npm registry, including @0xpolygonid/js-sdk, @iden3/js-iden3-core, and ethers. These are well-known packages in the self-sovereign identity ecosystem.
[PROMPT_INJECTION]: The SKILL.md and reference files contain instructions for the agent to check identity status before execution and to stop on script failure. These are interpreted as functional guardrails to ensure proper operation rather than malicious injection attempts.

verified-agent-identity