cc-prompt-builder
Pass
Audited by Gen Agent Trust Hub on May 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or unauthorized data exfiltration attempts were detected. The skill's behavior is consistent with its stated purpose of prompt engineering for developer tools.
- [DATA_EXPOSURE]: The instructions direct the agent to read project files such as source code, configuration files (package.json, tsconfig), and environment examples (.env.example). This is standard practice for providing necessary context to an LLM to generate accurate code-related prompts.
- [INDIRECT_PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection because it reads untrusted project files to build a prompt.
- Ingestion points: Reads project files, configuration files, and CI workflows.
- Boundary markers: None specified in the instructions.
- Capability inventory: File reading (via standard agent tools) and file writing (to save the generated prompt as markdown).
- Sanitization: No specific sanitization or escaping of the ingested file content is mentioned.
- Risk: Malicious instructions embedded in a project file could potentially influence the content of the generated prompt. However, as the output is a static document intended for user review and subsequent use with an external CLI tool, the risk remains low.
Audit Metadata