The Agent Skills Directory

[SAFE]: The skill operates within a restricted scope, focusing on project-level documentation. It does not contain instructions for command execution, privilege escalation, or data exfiltration. All file operations are limited to reading and writing common documentation files like PRODUCT.md.\n- [INDIRECT_PROMPT_INJECTION]: The skill possesses a common surface for indirect prompt injection because it incorporates the contents of existing repository files into its context.\n
Ingestion points: Reads PRODUCT.md, GLOSSARY.md, and docs/decisions/README.md.\n
Boundary markers: The instructions do not define specific delimiters or instructions for the agent to ignore potentially malicious content within these files.\n
Capability inventory: The skill's capabilities are limited to conversational interaction and updating the PRODUCT.md file. It does not utilize subprocesses, network tools, or administrative commands.\n
Sanitization: No explicit sanitization or validation of the ingested file content is performed.

define-product