The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes YouTube comments, which constitute untrusted data from external sources. Malicious actors could embed instructions within comments aimed at influencing the AI agent's behavior during analysis.
Ingestion points: Comments are retrieved from the YouTube platform via the Apify actor (SKILL.md).
Boundary markers: The skill does not explicitly define delimiters or warnings to ignore instructions embedded within the scraped comment text.
Capability inventory: The skill provides instructions for executing local scripts and shell commands (node, curl), which could be targeted by successful injection.
Sanitization: There is no evidence of filtering or sanitizing the ingested comment text before it is returned to the agent context.
[COMMAND_EXECUTION]: The skill instructions include examples of shell commands used to execute local scripts (node scripts/run_actor.js) and interact with remote APIs using curl.
[DYNAMIC_EXECUTION]: The skill supports a customMapFunction parameter, which accepts a JavaScript string to transform output data at runtime. This allows for the execution of user-provided code strings.

scraping-youtube-comments