The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted metadata from the YouTube Data API, including video titles, tags, and descriptions. This content is then used to generate a research report, creating a surface where malicious instructions embedded in video content could potentially influence the agent's behavior.
Ingestion points: Video metadata is fetched in scripts/research_topic.py, saved to a JSON file, and then read by the agent in Step 4 of the instructions.
Boundary markers: The instructions do not specify the use of delimiters or warnings to ignore instructions within the fetched metadata.
Capability inventory: The skill possesses the capability to execute shell commands and perform file operations.
Sanitization: The Python script provides a 200-character preview of video descriptions, but other metadata fields are not explicitly sanitized before being processed by the agent.
[COMMAND_EXECUTION]: The skill requires the execution of a bundled Python script to gather topic data.
Evidence: SKILL.md Step 3 provides the command template: YT_API_KEY=API_KEY python3 <skill-dir>/scripts/research_topic.py "TOPIC".
[EXTERNAL_DOWNLOADS]: The skill involves the installation of the official google-api-python-client package.
Evidence: The skill documentation notes a dependency on this official library for YouTube API interactions.

youtube-topic-researcher