I built a tool to edit my agents text. I need honest feedback
retained text
I got tired of trying to make my agents text sound human. No matter what skills or prompt I use, the text simply sounds distinctively AI. I tried the humanizer skill, which helps a bit, but still, anyone can spot the “AI style”. The problem is that for certain public facing text, like emails to clients, PR descriptions, commit messages, I do not want other people to recognize that it was written by AI. Anyway, I built this tool that combines an open source CLI that installs some hooks into certain commands (git commit, file write, etc…) with an MCP server that queues the messages. The agents are forced to send the messages via MCP and wait for my editing and approval. Now that I made it public, I’m still wondering whether it “matches” other people’s workflows. Do you also find yourself cycling through terminal windows, copy pasting the agents drafts into the prompt box and editing them there? Do you come back to find that your agent already committed and pushed code with a commit message that anyone can spot was written by AI?
document fetched 2026-10-05 00:19 · canonical https://reddit.com/r/AI_Agents/comments/1wxhaxj/i_built_a_tool_to_edit_my_agents_text_i_need