The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from the conversation context and codebase to generate summaries for external publication.
Ingestion points: SKILL.md instructions to use "current conversation context and codebase understanding".
Boundary markers: None present; there are no instructions to distinguish between data and potential instructions within the ingested context.
Capability inventory: The agent utilizes repository exploration tools for reading and the Linear API/CLI for writing to an external tracker.
Sanitization: No sanitization or verification of the external data is performed before it is published to the issue tracker.
[COMMAND_EXECUTION]: The instruction to "Explore the repo" necessitates the use of filesystem and codebase exploration tools to gather context.
[DATA_EXFILTRATION]: The skill transmits repository domain knowledge and conversation summaries to an external Linear project tracker. This is documented as the primary function of the skill and targets a well-known service.

to-prd