skills/tmrw-realityos/9pm-skill/9pm/Gen Agent Trust Hub

9pm

Pass

Audited by Gen Agent Trust Hub on Nov 9, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the deployment of user-provided applications, which involves running local build scripts (e.g., npm install, docker build). This creates a surface for indirect prompt injection if a project contains malicious configuration or lifecycle scripts.
  • Ingestion points: User-controlled project directories and files.
  • Boundary markers: The instructions explicitly warn the agent to "Only deploy directories you trust, the same way you'd trust running that project's build."
  • Capability inventory: The 9pm deploy command executes arbitrary build and lifecycle scripts defined within the project.
  • Sanitization: No sanitization is performed on project scripts; the skill relies on user-verified trust in the codebase being deployed.
  • [EXTERNAL_DOWNLOADS]: The skill directs the agent to download the ninepm CLI from the public npm registry and update the skill itself via GitHub. These are well-known services and the resources are owned by the verified vendor.
  • [COMMAND_EXECUTION]: The skill uses the 9pm CLI to perform deployment and management tasks. It includes specific guidance for configuring per-command approval in the agent environment to ensure destructive actions like app deletion require explicit user confirmation.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 9, 2026, 05:36 PM