lineup: Open-Source CLI That Catches Coding Agents Making Sneaky Edits
Several-Coyote3534 · reddit · 2026-11-19
A developer released lineup (v0.1.1), an open-source CLI for Claude Code and Codex that reports after every agent reply:
- Risky files (HIGH/MED/LOW) using fixed rules, not AI — "the agent doesn't grade its own homework"
- Out-of-scope edits: sensitive files your request never mentioned (.env, auth)
- Hidden changes: files changed but left out of the agent's reply
A lineup rollback command undoes individual files. No API key or account; nothing leaves the machine. Built from a Reddit thread collecting what edits users most want flagged. Only tested on Windows with Claude Code so far; the author seeks Mac/Linux and Codex testers.
More from coding & agent
- Flow Engineering Brings Agentic AI to Hardware Design; Rivian Engineers Grew 40 to 1,500 in 4 Months — PTrubey · 2026-11-19
- Linq raises $90M Series B led by a16z, has processed 1B+ agent messages — a16z · 2026-11-19
- Google Labs launches Stitch Loop, an always-on product agent for post-launch improvement — GoogleLabs · 2026-11-19
- Email APIs need finer scopes for agents: treat inbox as a tag, read by range — yenkel · 2026-11-19
- Same models, more bugs: Xint's harness finds 5 critical nginx bugs, 3 never found before — dyn___ · 2026-11-19
- How Much Inbox Access Should an Agent Get? Debating Fine-Grained Scopes — yenkel · 2026-11-19