The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) as it parses natural language conversations and screenshots to extract financial data. Maliciously crafted input in these sources could potentially attempt to influence the agent's behavior during the extraction or report generation process.
Ingestion points: Processes 'moments' (daily conversations) and payment screenshots from various platforms (WeChat, Alipay, etc.) as primary data sources.
Boundary markers: The provided documentation does not specify the use of clear delimiters or instructions to ignore embedded commands within the processed data.
Capability inventory: The skill writes structured financial data to ~/PhoenixClaw/Finance/ and interacts with the openclaw CLI to schedule recurring tasks.
Sanitization: No explicit sanitization or validation logic for the ingested natural language text is described beyond semantic pattern matching.
[COMMAND_EXECUTION]: The skill utilizes the openclaw command-line interface to manage scheduled tasks (cron jobs) for generating financial reports. While this is a documented feature of the platform and central to the skill's purpose, it represents a persistent execution capability that relies on the integrity of the agent's instructions.

phoenixclaw-ledger