The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process a 'source repo', including its internal prompts and instructions. This untrusted external data could contain malicious instructions meant to override the agent's behavior during the conversion process.
Ingestion points: Source repository files, prompts, and instructions mentioned in SKILL.md ('understand the ai repository').
Boundary markers: Absent. There are no instructions to delimit or ignore instructions found within the source repository.
Capability inventory: File system writing (creating .pi/ structure), CLI execution (pi print, pi terminal), and subagent orchestration.
Sanitization: Absent. The instructions state 'repo source code and prompts are only truth', indicating the agent will faithfully process all content found in the source.
[COMMAND_EXECUTION]: The skill requires the execution of CLI tools for auditing and verification.
Evidence: Instructions include pi print --approve and the use of the pi terminal tool for verification.
Context: The --approve flag suggests automated approval of loaded resources, which reduces manual oversight.
[EXTERNAL_DOWNLOADS]: The skill identifies specific external dependencies required for the conversion process.
Evidence: Lists npm:pi-subagents and npm:pi-hermes-memory as required packages.

ai-repo-to-pi-package