The Agent Skills Directory

[SAFE]: The skill serves as a structural guide for the agent to summarize conversation history into a 'workmap' format (ideas, projects, tasks, and questions). It does not include any scripts, network requests, or sensitive file access. All logic is contained within descriptive text instructions.
[NO_CODE]: The skill consists entirely of instructional markdown and platform-specific configuration files. There are no associated scripts, binaries, or external code dependencies, significantly reducing the potential attack surface.
[INDIRECT_PROMPT_INJECTION]: The skill operates on the active conversation thread as its data source. While this represents a standard ingestion surface for summarization tools, the skill instructions are strictly scoped to formatting and organizational logic. Crucially, the skill lacks access to any tools (such as shell execution or network calls) that could be leveraged for an attack even if malicious instructions were present in the processed data.

workmap