The Agent Skills Directory

[DATA_EXFILTRATION]: The skill communicates with an external endpoint at api.openclaw.ai to fetch and process data. Although this is consistent with its stated purpose, the domain is not categorized as a trusted or well-known service in the security framework.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it ingests and processes untrusted data from APIs and local JSON files. • Ingestion points: Data is loaded via the --input flag from local files (e.g., history.json) and retrieved from API responses at api.openclaw.ai. • Boundary markers: There are no defined delimiters or instructions to ignore embedded commands within the ingested data. • Capability inventory: The skill uses the claw CLI and network-capable libraries like requests, which could be exploited if malicious instructions are processed. • Sanitization: No input validation or data sanitization logic is documented to handle potentially malicious payloads in financial data strings.
[COMMAND_EXECUTION]: The skill is primarily accessed through the claw CLI, which executes system commands based on user-provided arguments such as symbols, intervals, and file paths.

market-data