scan-codebase-health
Pass
Audited by Gen Agent Trust Hub on Nov 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the codebase which can lead to indirect prompt injection if malicious instructions are embedded in files or configurations.
- Ingestion points: Project documentation (
docs/), source code files, and configuration files (defined byconfigPatterns) are read during Phases 1, 2, 5, 6, and 7 to verify counts, references, and patterns. - Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are utilized when reading content from the scanned project files.
- Capability inventory: The skill utilizes
spawn_agentto create sub-agents, executes local Python scripts (.claude/scripts/code_graph), and writes report files totmp/reports/. - Sanitization: There is no evidence of sanitization or filtering of external content extracted from the codebase before it is processed by the agent or included in the final report.
- [COMMAND_EXECUTION]: The skill executes a local Python utility to perform code graph queries.
- Evidence: In Phase 3 and Phase 4, the skill runs
python .claude/scripts/code_graph query importers_of <symbol> --json. While this targets an internal tool, the<symbol>or<file>arguments are derived from the codebase being scanned, presenting a potential (though localized) command injection surface if the codebase contains specifically crafted filenames or symbols.
Audit Metadata