security-hook-triggers-on-literal-exec-eval-tokens-in-prose

The security_reminder_hook regex catches literal exec( and eval( tokens anywhere in file content, including documentation and prose that merely describes the hook. This is a recurring meta-recursion: any handoff or SKILL.md that explains the hook will trigger it. Workaround: use descriptive language (‘shell execution call’, ‘dynamic evaluation’) instead of the literal token forms. Applies across all file types written via Edit/Write tools.

Related

• security-hook-fires-on-literal-exec-eval-tokens-in-prose
• security-hook-blocks-literal-exec-eval-in-prose
• security-hook-blocks-literal-exec-eval-tokens-in-prose
• security-hook-blocks-exec-eval-tokens-in-prose
• security-reminder-hook-blocks-literal-exec-eval-in-prose
• subagent-grep-operations-leave-temp-files-in-tmp
• fusion-invariants-lint-manual-no-drift-guard
• makefile-awk-help-regex-excludes-digit-targets