The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and analyzes tweet history from external Twitter profiles, which introduces a surface for indirect prompt injection if the tweets contain instructions targeted at the processing agent.
Ingestion points: Tweet text, engagement metrics, and author metadata are retrieved from Twitter via the apidojo/twitter-profile-scraper actor as specified in SKILL.md.
Boundary markers: The instructions do not define clear delimiters or include warnings to ignore instructions embedded within the scraped tweet content.
Capability inventory: The skill has the ability to execute shell commands using node and curl as shown in the examples in SKILL.md.
Sanitization: There is no description of filtering or sanitizing the retrieved tweet content before it is processed by the agent.
[DYNAMIC_EXECUTION]: The skill's input schema includes a customMapFunction parameter that accepts a JavaScript function as a string. This enables dynamic code execution within the remote environment to transform data objects at runtime.
[EXTERNAL_DOWNLOADS]: The skill communicates with the Apify platform (api.apify.com) to run the scraping actor and retrieve datasets. These interactions are consistent with the skill's purpose and utilize the author's own actor resources on a well-known service.
[COMMAND_EXECUTION]: Documentation in SKILL.md provides examples for executing system commands, including running a local Node.js script (node scripts/run_actor.js) and making network requests via curl to manage actor execution and output formatting.

analyzing-competitor-twitter-profile-content