The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The SKILL.md file contains a preamble section that attempts to execute local scripts (telemetry/version_check.py and telemetry/telemetry_init.py). These files are missing from the skill distribution, which prevents verification of their behavior and poses a risk of executing unreviewed code during skill initialization.
[PROMPT_INJECTION]: The skill processes untrusted external data from RSS feeds and transcript files. This data is interpolated directly into LLM prompts without boundary markers or instructions to ignore embedded commands, creating a surface for indirect prompt injection.
Ingestion points: podcast_pipeline.py (ingests content from RSS URLs and transcript files).
Boundary markers: Absent; the transcript text is passed directly into prompts for extraction and generation.
Capability inventory: The skill performs network requests (via requests and SDKs) and file system writes (to the ./output directory).
Sanitization: No filtering or sanitization of the input text is performed.
[EXTERNAL_DOWNLOADS]: The podcast_pipeline.py script automatically downloads audio files from remote URLs extracted from RSS feeds. While this is a core functionality, it involves downloading and processing files from arbitrary, potentially untrusted sources.

podcast-pipeline