The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The references/cli-actorization.md file contains a Dockerfile example that recommends installing a third-party utility (ubi) by downloading a bootstrap script from a remote GitHub repository and piping it directly into the shell. This pattern executes remote code without prior integrity verification or package manager safeguards.
Evidence: RUN curl --silent --location https://raw.githubusercontent.com/houseabsolute/ubi/master/bootstrap/bootstrap-ubi.sh | sh in references/cli-actorization.md.
[INDIRECT_PROMPT_INJECTION]: The skill documents and provides examples for building web scrapers and Actors that ingest data (such as page titles and web content) from external, untrusted websites, creating a surface for indirect prompt injection attacks.
Ingestion points: references/js-ts-actorization.md (via page.title()), references/python-actorization.md, and SKILL.md.
Boundary markers: The skill includes a dedicated "Security" section in SKILL.md that warns users to treat all crawled web content as untrusted input.
Capability inventory: The resulting Actors typically possess network access for crawling, file system access for storage, and interaction with the Apify platform via its CLI and SDK.
Sanitization: The skill provides specific guidance on sanitizing crawled data before passing it to shell commands, database queries, or template engines.

apify-actorization