codex-architect
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill employs the Bash tool to perform project discovery and analysis through ls, grep, and cat commands. These operations are scoped to the local repository for architectural evaluation.\n- [INDIRECT_PROMPT_INJECTION]: The skill has an inherent attack surface for indirect prompt injection as it processes file content from the local repository which could contain malicious instructions.\n
- Ingestion points: Repository files are read using cat and searched via grep as specified in the research steps in SKILL.md.\n
- Boundary markers: The skill lacks explicit delimiters to differentiate between code data and natural language instructions during analysis.\n
- Capability inventory: The agent is granted Bash and Write tool access, which could be targeted by successful injection.\n
- Sanitization: There is no evidence of filtering or escaping logic applied to the files read from the repository.
Audit Metadata