new-project-setup

Pass

Audited by Gen Agent Trust Hub on Sep 3, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes shell commands for repository setup, specifically git init to initialize the directory and init-cursor.sh to configure the Cursor editor environment.
  • [EXTERNAL_DOWNLOADS]: Fetches project license documentation from the FixMyBerlin organization's GitHub repository (FixMyBerlin/tilda-geo).
  • [REMOTE_CODE_EXECUTION]: Uses bunx to download and execute skill installation scripts from the official FixMyBerlin skills repository (FixMyBerlin/fixmyskills). These are recognized as vendor-owned resources.
  • [INDIRECT_PROMPT_INJECTION]: The skill accepts user-provided data such as project names and copyright holders which are used to configure the project metadata.
  • Ingestion points: User input for project name and copyright holder metadata in SKILL.md.
  • Boundary markers: None explicitly defined to separate user input from system instructions.
  • Capability inventory: Includes shell command execution (git, bunx) and file system write operations.
  • Sanitization: No explicit sanitization or validation of the user-provided strings is documented.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 3, 2026, 01:04 PM
Security Audit — agent-trust-hub — new-project-setup