The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources that could be used to influence the agent's behavior.
Ingestion points: The skill uses chrome-use to extract the full innerText of DeBank profile pages and executes a Python script that fetches JSON data from multiple external DeFi protocol APIs.
Boundary markers: Absent. There are no explicit delimiters or instructions provided to the agent to treat the extracted DOM text or API responses as untrusted data or to ignore embedded instructions.
Capability inventory: The agent can execute shell commands (via the chrome-use utility) and run local Python scripts.
Sanitization: No sanitization or validation of the fetched content is performed before the agent processes it for portfolio analysis.
[COMMAND_EXECUTION]: The skill executes shell commands to interact with a browser and run a local analysis script.
It invokes a utility located at /Users/engineer/.claude/skills/chrome-use/scripts/chrome-use to open browser tabs and evaluate JavaScript on external pages.
It executes a local Python script fetch_apys.py to aggregate and process financial data.
[EXTERNAL_DOWNLOADS]: The fetch_apys.py script performs network operations to fetch DeFi market data.
It connects to several well-known services including Morpho Blue (blue-api.morpho.org), DeFiLlama (yields.llama.fi), Ethena (ethena.fi), and Hyperliquid (api.hyperliquid.xyz). These are recognized as established and trusted financial data providers.

overview-crypto-portfolio