The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill attempts to read API keys from a specific local file path at ~/.secrets/aipodcasting/env and checks several environment variables (WIN_API_KEY, AIP_API_KEY, etc.) to authenticate requests to the backend media processing service in scripts/media_toolkit_lib/api.py.
[COMMAND_EXECUTION]: The script scripts/media_toolkit_lib/io.py executes a local binary located at ~/GitHub/scripts/bin/upload-media (or a path defined by the MEDIA_UPLOAD_BIN environment variable) using subprocess.run to handle file uploads.
[INDIRECT_PROMPT_INJECTION]: The transcribe command fetches transcript data from remote URLs and returns the text directly to the agent's context. This presents a surface for indirect prompt injection if the fetched transcript contains malicious instructions.
Ingestion points: The transcribe function in scripts/media_toolkit.py calls api_client.fetch_text() to retrieve content from transcript_url.
Boundary markers: The output is returned as raw text or within a JSON object without explicit boundary delimiters or warnings to ignore embedded instructions.
Capability inventory: The skill has the ability to execute local commands via subprocess.run and write files to the local system via write_json_file in scripts/media_toolkit_lib/io.py.
Sanitization: No sanitization or filtering is performed on the fetched transcript text before it is returned to the agent context.

media-toolkit