c4-architecture-c4-architecture
Pass
Audited by Gen Agent Trust Hub on Jul 7, 2026
Risk Level: SAFEPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes untrusted content from the repository being analyzed.
- Ingestion points: All files in the repository subdirectories identified during Phase 1.1 and 1.2 are passed to subagents for analysis.
- Boundary markers: The prompts for subagents (e.g., c4-architecture::c4-code) do not include explicit delimiters or instructions to ignore embedded commands or instructions within the source code.
- Capability inventory: The skill performs extensive file system read operations and writes generated documentation to the C4-Documentation/ directory.
- Sanitization: The skill does not implement sanitization or validation of the ingested code before it is processed by the AI subagents.
- [DATA_EXFILTRATION]: The skill's primary function involves a deep scan of the codebase, which creates a risk of exposing sensitive information such as credentials or private keys in the resulting documentation.
- Evidence: The instructions in Phase 1.1 and 1.2 command the agent to process every subdirectory and all code elements, which may include configuration files or environment secrets.
- Mitigation: Although the skill provides an exclude_patterns configuration option, the default behavior and the requirement for manual configuration mean that sensitive data could be inadvertently included in the generated architectural documentation if the user does not apply proper exclusions.
Audit Metadata