AGENT LAB: SKILLS

apify-audience-analysis

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
  • EXTERNAL_DOWNLOADS (MEDIUM): The skill instructs the user to install the '@apify/mcpc' package globally. This is an external dependency that is not managed by the skill itself.
  • COMMAND_EXECUTION (MEDIUM): The workflow executes a local script at '${CLAUDE_PLUGIN_ROOT}/reference/scripts/run_actor.js'. Because the script body is not provided in the skill definition, its execution logic and safety cannot be verified.
  • CREDENTIALS_UNSAFE (LOW): The skill processes 'APIFY_TOKEN' from a local '.env' file. While environment variable usage is standard, extracting them via shell commands (grep/xargs) increases visibility of the secret in process lists.
  • PROMPT_INJECTION (LOW): This skill has an indirect prompt injection surface. It fetches dynamic schemas and data from the Apify platform (Step 2 and Step 5). Malicious content scraped from social media or provided in actor descriptions could influence the agent's behavior or summaries without proper sanitization.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 04:40 PM