codex-hygiene-skill
Warn
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill installation involves cloning a repository from an unverified GitHub account (sunflower-of-parchman/codex-hygiene).
- [COMMAND_EXECUTION]: The skill requires the execution of downloaded shell scripts (e.g., measure_codex_context.sh) to perform its primary function.
- [DATA_EXFILTRATION]: The skill accesses sensitive local application data, including configuration files (
/.codex/config.toml) and telemetry databases (/.codex/telemetry.db), which contain history of model interactions. - [INDIRECT_PROMPT_INJECTION]: By processing telemetry databases that store past model interactions, the skill creates an attack surface for indirect prompt injection if those logs contain malicious instructions (Category 8c).
Audit Metadata