apify-actor-development
Pass
Audited by Gen Agent Trust Hub on Sep 24, 2026
Risk Level: SAFEPROMPT_INJECTIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill's YAML frontmatter and body contain instructions using the "Important:" override marker to compel the agent to populate specific metadata fields (generatedBy) with details about its own model and tool identity.
- [INDIRECT_PROMPT_INJECTION]: As a framework for building web crawlers and automation agents, the skill creates a surface for indirect prompt injection from untrusted web content processed at runtime.
- Ingestion points: Actor source code (src/main.js/ts/py) is designed to ingest and process data from external websites.
- Boundary markers: The instructions explicitly advise treating crawled content as untrusted and recommend strict validation and ignore-markers.
- Capability inventory: The skill enables project bootstrap, dependency installation, local execution, and deployment via the apify CLI and standard package managers.
- Sanitization: The documentation mandates input sanitization, prohibits the use of eval() or shell execution on external data, and requires using specialized logging libraries for credential censoring.
- [COMMAND_EXECUTION]: The skill utilizes legitimate developer tools including the apify CLI for project orchestration and npm or pip for managing runtime dependencies.
- [EXTERNAL_DOWNLOADS]: The skill references documentation and resource files from official Apify and Crawlee domains and their public GitHub repositories.
Audit Metadata