The Agent Skills Directory

[SAFE]: The skill implements a robust 'Documentation-Only Scope' policy that explicitly forbids the agent from editing source code, running build/migration commands, or modifying system configurations. This acts as a strong internal safeguard against unintended implementation actions.
[SAFE]: File system access is restricted to reading relevant project documentation (such as product briefs and research files) and writing new PRD documents to the 'docs/' directory. No access to sensitive system files or credentials was detected.
[SAFE]: The skill uses established internal orchestration patterns, calling other specific skills like 'create-doc' and 'pm-checklist' for its functional logic without introducing external code dependencies or remote script execution.
[SAFE]: No obfuscation, data exfiltration patterns, or prompt injection vulnerabilities were identified in the skill's instructions or metadata.

create-prd