screenplay

Pass

Audited by Gen Agent Trust Hub on Nov 4, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill relies on the execution of the npx agent-screenplay command to check, record, and render video scenarios. This involves calling an external CLI tool during the skill's operation.
  • [EXTERNAL_DOWNLOADS]: The skill downloads a JSON validation schema from the author's GitHub repository (https://raw.githubusercontent.com/gmpetrov/agent-screenplay/main/schema/scenario.schema.json) and fetches the agent-screenplay package from NPM via npx.
  • [REMOTE_CODE_EXECUTION]: The use of npx agent-screenplay constitutes the download and execution of code from a remote registry at runtime. While these are vendor-owned resources, they represent a remote code execution vector.
  • [INDIRECT_PROMPT_INJECTION]: The skill instructions create a vulnerability surface for indirect prompt injection: 1. Ingestion points: The agent is tasked with "exploring the app" and "reading its code" to identify UI elements (SKILL.md, Step 2). 2. Boundary markers: No explicit delimiters or instructions are provided to ignore malicious content within the target application's code or UI. 3. Capability inventory: The agent can execute shell commands, write files to the local system, and access network resources. 4. Sanitization: There is no requirement to sanitize or filter data extracted from target applications before it is incorporated into scenario generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 4, 2026, 08:04 PM