The Agent Skills Directory

[SAFE]: No malicious patterns or security risks detected. The skill focuses on structured documentation and requirement gathering.
[DATA_EXPOSURE]: The skill interacts with the local workspace to identify feature placement and project structure but does not access sensitive system files (e.g., SSH keys, credentials) or perform network exfiltration. It explicitly forbids the inclusion of external URLs in its output.
[PROMPT_INJECTION]: The instructions do not contain markers for overriding agent behavior, bypassing safety guidelines, or extracting system prompts.
[COMMAND_EXECUTION]: The skill defines a manual workflow for planning and documentation. It does not invoke shell commands, execute scripts, or perform privilege escalation.
[REMOTE_CODE_EXECUTION]: There are no remote code patterns or external script downloads. All referenced templates are local to the skill's own directory structure.

spec-architect