The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from the rorkai/app-store-connect-cli-skills repository and the output of the asc CLI tool. This creates a surface for indirect prompt injection where malicious content in the repository or CLI help could influence the agent. However, the risk is minimized by instructions to validate changes against schemas and help outputs, and the ingestion targets are vendor-owned resources.
Ingestion points: rorkai/app-store-connect-cli-skills repository and asc CLI help/schema outputs.
Boundary markers: The skill instructs the agent to "record exact help or runtime evidence for every edit" and follow AGENTS.md protocols.
Capability inventory: The skill can read file systems, perform network reads (git checkout), write to files (proven corrections), and create pull requests.
Sanitization: Instructions require validating command examples against current help and relevant schemas before application.
[EXTERNAL_DOWNLOADS]: The skill requires resolving a checkout of the rorkai/app-store-connect-cli-skills GitHub repository. This is a vendor-owned resource central to the skill's purpose of synchronizing documentation.
[COMMAND_EXECUTION]: The instructions involve executing the asc CLI utility with commands like --help, search, schema, and capabilities to verify the current state of the CLI surface. Execution and runtime verification are gated by explicit authorization requirements.

sync-asc-skills