The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from Instagram (captions, user handles) and incorporates it into reports and outreach recommendations. While it outlines steps to filter and rank this data, it lacks explicit boundary markers or instructions to treat the ingested text as untrusted data, which could lead to the agent following instructions embedded in scraped captions.
Ingestion points: Scraped Instagram data (caption, ownerUsername) from apidojo~instagram-location-scraper (SKILL.md).
Boundary markers: Absent. The agent is instructed to extract and rank content directly.
Capability inventory: Executes shell commands via node scripts/run_actor.js and curl (SKILL.md).
Sanitization: Absent. Captions are extracted and displayed in reports verbatim.
[DYNAMIC_EXECUTION]: The skill's input schema includes a customMapFunction parameter described as a "JavaScript function to transform each output object." If this function is executed by the agent or the underlying actor using untrusted input, it presents a risk of arbitrary code execution.
[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute shell commands using node and curl to interact with the Apify API. While these are intended for the skill's primary purpose, they represent a high-privilege capability that should be monitored.

scraping-instagram-location-content