The Agent Skills Directory

[PROMPT_INJECTION]: The SKILL.md file contains a 'Need full version? (Optional)' section that includes a block of text specifically designed to be read and executed as instructions by the AI agent. These instructions direct the agent to purchase and install an upgraded version of the skill via a third-party payment platform.
[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute the command npx -y @alipay/agent-payment@latest install. This command downloads and executes code from a remote registry to facilitate payment processing. While the provider is a well-known financial service, the skill proactively initiates external software installation.
[INDIRECT_PROMPT_INJECTION]: The skill processes external text and JSON files provided by the user (ingested via scripts/run.mjs using fs.readFileSync).
Ingestion points: User-specified files provided to the --input argument in scripts/run.mjs.
Boundary markers: None identified; the engine treats all input text as data to be parsed into a table.
Capability inventory: The skill has Read and Bash permissions; it executes local scripts to perform calculations.
Sanitization: The engine (scripts/engine/progress-payment-check.js) parses strings into numbers for arithmetic verification but does not sanitize content against potential downstream prompt injection if the results are interpreted by the agent.

progress-payment-check-free