The Agent Skills Directory

[SAFE]: The skill is designed to recap the current conversation context and verify task states using read-only tools like ls, git log, and file reading. It does not perform any write operations or external network requests during execution.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it processes untrusted user input from previous conversation rounds. This is a primary function of the skill. * Ingestion points: The skill re-reads the last N exchanges from the active conversation context as defined in SKILL.md. * Boundary markers: No specific delimiters or boundary markers are instructed for the agent to use when processing these exchanges. * Capability inventory: The skill allows the agent to use ls, git log, and read files to verify the state of the workspace as defined in SKILL.md. * Sanitization: There is no explicit sanitization or filtering of the ingested conversation history before it is summarized.

conclude-rounds