The Agent Skills Directory

[SAFE]: No malicious patterns or security violations were detected. The skill focuses on legitimate brainstorming and project management tasks.
[COMMAND_EXECUTION]: The skill instructs the agent to use local CLI tools such as git and flow-kanban to retrieve project status and logs. This is standard behavior for a development assistant.
[DATA_EXFILTRATION]: The agent reads local project context from MEMORY.md and git log. There are no network operations or patterns suggesting that this data is transmitted to external servers.
[PROMPT_INJECTION]: No direct prompt injection or bypass techniques are present. The skill possesses an Indirect Prompt Injection surface (Category 8) due to its reliance on project data:
Ingestion points: git log, MEMORY.md, and tool output from flow-kanban (SKILL.md).
Boundary markers: Not specified; the skill does not define delimiters to separate project context from instructions.
Capability inventory: Can create tickets via flow-kanban create or hand over to grill-me (SKILL.md).
Sanitization: No explicit validation or filtering is performed on data retrieved from the project environment.

brainstorming