The Agent Skills Directory

[SAFE]: The skill provides documentation and code examples for interacting with the official Trust Wallet API at tws.trustwallet.com. The authentication mechanism (HMAC-SHA256) is correctly described and follows security best practices for API communication.
[SAFE]: The skill includes a dedicated security reference (references/security.md) that allows the agent to perform risk assessments, including honeypot detection, smart contract audit verification, and address blacklisting checks before suggesting transactions to the user.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external API endpoints (e.g., token names, symbols, and security warnings from tws.trustwallet.com).
Ingestion points: Data enters the context via GET /v1/search/assets, GET /v2/coinstatus/{assetId}, and POST /amber-api/v1/route in the various reference files.
Boundary markers: None are explicitly defined in the skill instructions to separate API responses from internal agent logic.
Capability inventory: The skill facilitates the retrieval of transaction data (evmTx.data, data) for blockchain swaps through the POST /amber-api/v1/route/step endpoint.
Sanitization: The skill provides security metadata (risk levels, honeypot status) to help the agent evaluate the safety of the tokens being queried, though it does not provide specific prompt-level sanitization for the API's JSON output.

trust-wallet-api