apify-easy-competitive-intelligence
Pass
Audited by Gen Agent Trust Hub on Aug 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructions direct the agent to execute shell commands using the
apifyCLI for actor invocation, dataset retrieval, and schema discovery. It also includes the use ofjqfor JSON processing and thesourcecommand for environment variable management. - [REMOTE_CODE_EXECUTION]: The skill utilizes
python3 -cto execute dynamically generated Python logic for data aggregation, extraction, and transformation of scraped results. Additionally, it calls remote Apify actors, which execute code on the Apify platform. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted data from multiple external platforms, including product reviews, social media posts, and job listings.
- Ingestion points: Data retrieved from various Apify actors (e.g., G2, Capterra, Reddit) is saved to
/tmp/results.jsonand read into the agent's context for analysis. - Boundary markers: The instructions lack explicit boundary markers or directives for the agent to ignore potentially malicious instructions embedded within the scraped data.
- Capability inventory: The environment provides shell execution, Python script execution, and local file system access.
- Sanitization: There are no instructions for sanitizing or filtering external content before it is processed by the agent.
Audit Metadata