cursor-hello-world
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill establishes a prompt injection surface by guiding the agent to process external instructions through Cursor's AI features.
- Ingestion points: Sample prompts in SKILL.md (e.g., instructions for formatCurrency or task manager) intended for AI processing.
- Boundary markers: Explicit instructions to 'inspect every generated diff before accepting it' and to 'reject or revert unexpected edits' are present in the instructions.
- Capability inventory: The skill possesses the ability to create/edit files and execute bash commands (npm, tsx) within the environment.
- Sanitization: No programmatic sanitization; the skill relies on manual user review as a security boundary.
- [COMMAND_EXECUTION]: The skill executes shell commands to initialize a local project and install development dependencies.
- Evidence: npm init -y, npm install typescript tsx @types/node, and npx tsc --init.
- Context: These commands are used to set up a standard TypeScript development environment for a tutorial and align with the primary purpose of the skill.
Audit Metadata