The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it ingests and processes text output (readable_text) generated by a remote backend agent to drive user interactions and workflow decisions.
Ingestion points: The readable_text field returned by the get-thread command in SKILL.md is used to determine the next steps and generate questions for the user.
Boundary markers: There are no explicit instructions or delimiters used to isolate the server-provided text from the agent's internal logic.
Capability inventory: The skill has the capability to execute shell commands (pippit-tool-cli) and perform file system writes (download-result).
Sanitization: The instructions include logic for filtering and cleaning workflow options based on the drama creation stage, but do not specify security-focused sanitization for the raw text content.
[EXTERNAL_DOWNLOADS]: The skill provides instructions to install its required tooling via npx @pippit-dev/cli@latest install. This package is a vendor-owned resource hosted on the well-known NPM registry, representing standard installation behavior.
[COMMAND_EXECUTION]: The skill functions by orchestrating the pippit-tool-cli to submit drama creation runs, upload reference files, and download generated assets like scripts and videos to the local filesystem.

xyq-short-drama-skill