The Agent Skills Directory

[SAFE]: The skill is restricted to generating markdown documentation. It does not perform any network operations, access sensitive credentials, or execute arbitrary shell commands.
[SAFE]: Explicit 'HARD-GATE' and 'Strict no-implementation' instructions are present to prevent the agent from crossing into code generation or implementation, ensuring the skill remains focused on architectural planning.
[SAFE]: File system activity is limited to reading from and writing to the project's docs/ directory. Path patterns are well-defined and follow standard documentation practices.
[SAFE]: Although the skill ingests untrusted requirements data, it lacks the exploitable capabilities (such as code execution or network exfiltration) required to facilitate a high-severity indirect prompt injection attack. Output is limited to markdown snapshots.
[SAFE]: The metadata and references to external design methodologies are used neutrally for instructional purposes and do not involve downloading or executing external scripts.

design-solution