14days-build-claude-code-cli
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions to clone an external repository from GitHub (https://github.com/bozhouDev/14days-build-claude-code-cli.git) as part of its tutorial walkthrough.
- [COMMAND_EXECUTION]: Contains Python code implementation examples for a bash tool using subprocess.run(shell=True) to execute shell commands. The tutorial teaches best practices for mitigating risk, including checking for dangerous command patterns (e.g., sudo, rm -rf) and requiring explicit user permission before execution.
- [PROMPT_INJECTION]: As an educational guide for building agents, the skill highlights the attack surface for indirect prompt injection. It includes a mandatory evidence chain for this risk: (1) Ingestion points: The read_file_tool in SKILL.md reads local files into the agent context. (2) Boundary markers: Implementations in SKILL.md include checks to ensure paths remain within the current working directory. (3) Capability inventory: The agent architecture includes subprocess calls (bash_tool) and file system writes (write_file_tool). (4) Sanitization: The tutorial mandates user confirmation prompts for all file edits and dangerous shell commands to prevent unauthorized actions.
Audit Metadata