The lesson
Write a security guide as a run of named attacks, each with a short demo a reader can reproduce and a plain mitigation right after it.
An intro frames AI editors as 'execution environments' and says the goal is to treat them as 'attack surfaces to defend rather than trusted copilots', with a byline and a sticky table of contents. Sections such as 'Prompt Injection in Code Comments', 'Path Traversal in Claude Code (CVE-2025-54794)' and 'Remote Code Execution in Cursor via Config Override' each explain the flaw, show code or commands that trigger it, and close with mitigations. Later sections cover 'Misconfiguration as an Attack Vector', 'Mapping to MITRE ATT&CK' and 'Securing AI Code Editors', ending with a conclusion and a references list.
Don't copy Kodem's exploits, CVE write-ups or its framing of specific editors. Keep the attack, demo, mitigation rhythm per section, the framework mapping, and the references list that backs the claims.
Tell us what you want to make. We’ll talk through how to put these ideas to work for your brand.