The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The documentation in SKILL.md provides instructions for a manual upgrade process that involves the agent executing @alipay/agent-payment via the npx utility. This allows for the download and execution of a tool from a well-known payment provider to facilitate commercial features.
[INDIRECT_PROMPT_INJECTION]: The skill's analysis engine processes untrusted table data provided by the user, creating a vulnerability surface for embedded instructions.
Ingestion points: Data is ingested through the text field in the input JSON or via a file read in scripts/run.mjs.
Boundary markers: No delimiters or defensive instructions are used to distinguish user data from agent commands during the parsing process.
Capability inventory: The skill is authorized to use Read and Bash tools, which could be misused if instructions in the data were followed by the agent.
Sanitization: Input content is parsed numerically and split by whitespace/tabs without sanitization of the text itself.
[EXTERNAL_DOWNLOADS]: The skill references external resources for its upgrade path, including the @alipay/agent-payment package and official vendor domains like skillpay.alipay.com and chenqg618.github.io. These are documented as part of the intended commercial workflow of the skill.

restaurant-food-cost-check-free