The Agent Skills Directory

[Indirect Prompt Injection] (SAFE): The skill ingests untrusted data from external sources such as git logs and project file structures during its context gathering phase. While this provides an attack surface where a malicious user could embed instructions in commit messages to influence the brainstorming process, this risk is mitigated by the skill's design, which requires constant user validation and incremental step-by-step confirmation. Evidence: Phase 1 context gathering reads project structure and git logs; Phase 4 requires user confirmation for every 200-300 word section.
[Command Execution] (SAFE): The skill utilizes git commands (git log, git commit) and file writing operations to docs/plans/. These actions are necessary for the skill's functionality as a design documentation tool and are performed in a transparent manner.
[Data Exposure & Exfiltration] (SAFE): No unauthorized network operations or exfiltration of sensitive files were detected. Interactions are limited to local project files and trusted internal skill integrations.

brainstorm