codex-fable5
Fail
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the user to download and install external code from an unverified GitHub repository (
baskduf/FableCodex) using thecodex plugin marketplace addcommand. This source is not associated with a known trusted organization. - [REMOTE_CODE_EXECUTION]: The instructions direct users to execute binaries downloaded from a remote source. It provides specific commands to modify the shell environment PATH to include these third-party binaries and subsequently run them:
export PATH="$PWD/plugins/codex-fable5/bin:$PATH"andcodex-fable5 status. - [COMMAND_EXECUTION]: The skill contains instructions for the agent to execute shell commands for plugin management and workflow automation (e.g.,
codex plugin add,codex-fable5 goals next), which triggers the execution of external scripts not included in the skill's distribution. - [PROMPT_INJECTION]: The skill is designed to ingest and process untrusted external data in the form of AI prompts from other providers ("Convert this Claude/Fable prompt into Codex project rules"). This creates an ingestion point for indirect prompt injection where malicious instructions could be embedded in the processed prompts to influence agent behavior.
Recommendations
- AI detected serious security threats
Audit Metadata