The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Task tool to dispatch a specialized search-conversations subagent and provides documentation for direct access to MCP tools (mcp__plugin_episodic-memory_episodic-memory__search and mcp__plugin_episodic-memory_episodic-memory__read). These tools are used to query and read local conversation archives.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks because it processes data from past conversations which may contain untrusted content.
Ingestion points: The read tool (described in MCP-TOOLS.md) ingests raw conversation data from .jsonl files located in the ~/.config/superpowers/conversation-archive/ directory.
Boundary markers: The instructions do not define boundary markers or explicit warnings for the agent to ignore instructions found within the retrieved historical data.
Capability inventory: The skill facilitates complex task dispatching via the Task tool and subagent system, providing a path for retrieved instructions to influence agent behavior.
Sanitization: There is no mention of sanitization or validation logic for the content retrieved from historical archives.

remembering-conversations