The Agent Skills Directory

[SAFE]: The skill implements a structured requirement gathering process consisting of an intent capture phase and a stress-test phase, which encourages thorough validation of user inputs.
[SAFE]: An information barrier is enforced by explicitly instructing the agent to ignore codebase, architecture, and design documents, preventing the agent from making assumptions based on existing structures.
[SAFE]: File system interactions are limited to reading a project glossary (docs/glossary.md) for terminology alignment and writing a summary brief (.brief.md) upon user request, which are typical tasks for a development assistant.
[SAFE]: No evidence of prompt injection, data exfiltration, obfuscation, or remote code execution was found in the skill instructions.

discovery