scaffold-exercises
Pass
Audited by Gen Agent Trust Hub on Aug 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands such as
mkdir -pandgit mvto create and reorganize directory structures. It also runspnpm ai-hero-cli internal lintto validate the created files. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided exercise plans to define the file system structure and content. This creates a surface for indirect prompt injection where malicious input in a plan could attempt to cause command injection or influence agent behavior.
- Ingestion points: User-provided plans containing section and exercise names are parsed in the 'Parse the plan' step (SKILL.md).
- Boundary markers: The skill does not provide explicit boundary markers or instructions to the agent to disregard instructions potentially embedded within the plan text.
- Capability inventory: The skill utilizes file system creation (
mkdir), version control operations (git mv), and execution of a local CLI tool (pnpm ai-hero-cli). - Sanitization: Although the skill suggests a dash-case naming convention, it does not provide explicit instructions to sanitize or escape user-controlled strings before they are used in shell commands.
Audit Metadata