record-promo

Warn

Audited by Gen Agent Trust Hub on Nov 6, 2026

Risk Level: MEDIUMDYNAMIC_EXECUTIONCOMMAND_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The skill is designed to analyze a game's source code and dynamically generate a new script (scripts/capture-promo.mjs) tailored to that game. This generated script is subsequently executed using Node.js to perform automated gameplay capture.
  • [COMMAND_EXECUTION]: The skill performs multiple shell command executions, including running a development server via npm, executing the generated capture script with node, running conversion logic with bash, and using ffmpeg for video manipulation.
  • [PRIVILEGE_ESCALATION]: The skill explicitly instructs the agent to modify file permissions using chmod +x on a shell script (convert-highfps.sh) that it copies into the project directory.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the user's game project which could contain instructions meant to subvert the agent's behavior during the script generation phase.
  • Ingestion points: The agent reads src/scenes/GameScene.js, src/core/EventBus.js, src/core/Constants.js, and src/main.js from the provided game path.
  • Boundary markers: There are no boundary markers or instructions to ignore potential commands embedded in the source files being analyzed.
  • Capability inventory: The skill possesses capabilities for file system modification, permission changes (chmod), and arbitrary command execution (node, bash).
  • Sanitization: No sanitization or validation is applied to the content of the source files before they are used to influence the logic of the generated scripts.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Nov 6, 2026, 04:39 AM