Off the Wall: recall your past corrections before Codex answers

[toranalytics]

I built Off the Wall because I kept repeating the same corrections to my coding agent. It works with Codex or Claude Code, whichever you use. I am the author; the project is MIT-licensed.

GitHub: https://github.com/toranalytics/off-the-wall

It keeps corrections in your own words and quotes relevant past corrections before the next answer, rather than only turning them into summary rules. otw report also shows which kinds of corrections you repeat. --seed optionally starts recall from existing history.

Update: I found automated session messages mixed into my local report input. I have removed the earlier personal counts and report image while I audit the input filter. The report should not be read as evidence of reduced mistakes or repeated corrections.

Codex installation:

codex plugin marketplace add toranalytics/off-the-wall
codex plugin add off-the-wall@off-the-wall

Open /hooks once and trust the plugin hooks. Requires Node 18 or newer.

No separate API key is needed. The judge sends excerpts through your own Claude/Codex CLI to its provider; stored corrections remain on your machine.

If you try it, I would value feedback on missed corrections or recalled quotes that are not relevant. Counts-only reports are welcome; there is no need to share private conversations.

Plain-language overview of what it does and does not do, with the test details at the end: https://toranalytics.github.io/en/otw-what-it-does/