The Agent Skills Directory

[DATA_EXPOSURE_AND_EXFILTRATION]: The skill scans codebase metadata, including configuration files and environment variables, to document architecture and business rules. While this involves accessing sensitive configuration areas, the operations are limited to reading local files for the purpose of generating local documentation. No network tools or external communication patterns were identified.
[INDIRECT_PROMPT_INJECTION]: As the skill is designed to ingest and process source code from a directory, it has an inherent surface for indirect prompt injection via malicious code comments. The instructions attempt to mitigate this by directing the agent to summarize intent and contracts rather than copy-pasting raw implementation, though it lacks formal delimiters for the ingested code blocks.
[COMMAND_EXECUTION]: The skill uses file-system reconnaissance to build a mental model of the codebase. It does not invoke arbitrary shell commands or system-level utilities beyond standard file reading and writing within the project scope.

yds-spec-doc