scaffold-exercises
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill performs several local command-line operations to scaffold the directory structure. This includes the use of
mkdir -pto create nested directories,git commitandgit mvfor repository management, andpnpm ai-hero-cli internal lintto validate the created structure. These commands are executed directly in the project environment.\n- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it parses exercise plans from the context to determine names and content for files and directories.\n - Ingestion points: The exercise 'plan' described in the workflow section of SKILL.md.\n
- Boundary markers: The instructions do not define any delimiters to separate the plan content from the agent's instructions, nor do they instruct the agent to ignore embedded commands.\n
- Capability inventory: The skill has the capability to write to the file system, move files via version control, and execute a local CLI tool via pnpm.\n
- Sanitization: There is no explicit sanitization or validation of the strings extracted from the plan before they are used in shell commands, other than a suggestion to use dash-case naming.
Audit Metadata