AI coding tools keep leaking conversation history into commit messages and comments
recro69 · reddit · 2026-09-15
The author noticed AI coding agents leak conversation context into the codebase: after asking Claude to remove exponential backoff logic, the commit message read "Add logic (without exponential backoff)" — even though nobody ever asked for backoff; Claude added it, got corrected, and that correction became permanent project history.
Similar traces show up in comments, test names, filenames, and PR descriptions, like "as discussed" or ", per your feedback". The odd part: the code itself is usually fine — Claude just fails to switch audiences, still talking to the chat partner instead of a reader six months later with zero context. The author suggests we need a check for anything an AI agent creates that's supposed to stand on its own.
More from coding & agent
- GPT-6 Astra demos: playable games, robot demos and 3D pipelines built in hours — socialwithaayan · 2026-09-16
- Proposal: hand predictable UI workflows to a local runner to cut computer-use agent costs — Iinganouo · 2026-09-16
- Tutorial: Tracing agent invocations and LLM calls with OpenTelemetry — dl_weekly · 2026-09-16
- Atria Dawn Preview open-sourced: 744B MoE agent model tops BrowseComp at 92.5 — alexcovo_eth · 2026-09-16
- LangChain makes every Managed Deep Agent an MCP server callable by coding assistants — LangChain · 2026-09-16
- LlamaIndex on Building SDKs with Stainless, Whose Team Joined Anthropic — llama_index · 2026-09-16