apify-actor-development
Pass
Audited by Gen Agent Trust Hub on Oct 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill is a documentation-centric resource that focuses on secure development workflows. It includes specific sections on security best practices, such as sanitizing untrusted inputs and managing API tokens safely.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a framework for building actors that ingest external web data. It includes mandatory security requirements to validate and sanitize this untrusted input to prevent potential injection attacks from crawled content.
- Ingestion points: Untrusted web content scraped from external sites (SKILL.md).
- Boundary markers: Explicit warnings to treat all crawled content as untrusted (SKILL.md).
- Capability inventory: Network access for scraping and local storage operations (SKILL.md, references/dataset-schema.md).
- Sanitization: Mandatory instructions to sanitize crawled data and avoid execution of untrusted content (SKILL.md).
- [CREDENTIALS_UNSAFE]: Instructions explicitly warn against unsafe practices like passing secrets in command-line arguments or embedding them in source code, instead recommending environment variables and interactive login.
- [EXTERNAL_DOWNLOADS]: Recommends the installation of the Apify CLI via verified package managers (NPM, Homebrew) and provides links to official documentation on trusted domains (apify.com, crawlee.dev, github.com).
Audit Metadata