The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Python scripts using the shell: python scripts/quality_gates/four_us_score.py and python scripts/quality_gates/banned_word_check.py. These scripts are part of the Kai installation environment and are used to validate generated assets.
[DYNAMIC_EXECUTION]: The skill invokes Python scripts at runtime to perform quality gate checks on episode outlines, show descriptions, and pitch emails.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data from RSS feeds, episode URLs, and transcripts provided by the user.
Ingestion points: RSS metadata, episode URLs, and transcript text are loaded into the agent's context and logged in workspace/podcast/_ingest-log.md.
Boundary markers: The instructions do not specify the use of delimiters (like XML tags or triple quotes) or specific "ignore" instructions when reading external transcript data.
Capability inventory: The agent has the capability to write files to the workspace/ directory and execute local Python scripts via subprocess calls.
Sanitization: There is no mention of sanitizing or filtering the content of transcripts or RSS feeds to prevent embedded instructions from overriding the agent's current task.

kai-podcast