project-skill-audit
Pass
Audited by Gen Agent Trust Hub on Jul 9, 2026
Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DATA_EXFILTRATION]: The skill accesses potentially sensitive agent state and history. The instructions direct the agent to read files within
~/.codex/memories/and~/.codex/sessions/. These files contain records of past interactions and agent memory, which may include sensitive information or credentials provided in previous sessions. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. By analyzing past sessions and project documentation (e.g.,
README.md,AGENTS.md) to generate recommendations, the agent may process malicious instructions embedded in those files by an external actor or from a previous untrusted interaction. This could influence the agent's logic or the content of the recommended skills. - [COMMAND_EXECUTION]: The workflow involves executing local search commands. The skill explicitly instructs the agent to use
rg(ripgrep) to search through the memory index and session files. While ripgrep is a standard utility, its execution is gated by paths derived from environment variables or user-controlled project structures.
Audit Metadata