The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and interpret source code, tests, and configuration files. If an attacker embeds malicious instructions within code comments, test assertions, or strings in the target codebase, the agent might inadvertently execute those instructions while performing its extraction logic.
Ingestion points: Reads files found by find commands in PHASES 1 through 3 using cat.
Boundary markers: The skill instructions do not specify any delimiters or warnings to the agent to disregard instructions found within the processed files.
Capability inventory: The skill has filesystem read/write access and the ability to execute shell commands.
Sanitization: There is no evidence of filtering or sanitizing the content of the read files before they are processed by the agent's logic.
[COMMAND_EXECUTION]: The skill uses local shell commands including find, cat, and head to perform discovery and inspection of the project structure and source code. These operations are localized to the current directory and are used for documentation purposes.

business-logic-extractor