Caprock Files

Added

• Codex sessions are on the dashboard. A machine running Claude Code, OpenCode and OpenAI Codex had its spend split three ways, each tool knowing only its own share. Codex sessions, turns, tool calls and cost now sit on the same screens as the others, filterable by agent. Nothing to install: Codex writes one transcript per session and Caprock reads it, the way it already reads OpenCode's database.

Most Codex sessions cannot be priced, and the dashboard says so instead of guessing. Only 4 of 100 real transcripts named the model they ran on — the other 96 carry real token counts and no model at all, which is 83% of tokens. Those turns keep their tokens and carry no cost, and caprock status reports how many, so a partial total reads as partial. Inventing the model from the originator would have put a confident dollar figure on screen derived from nothing.

Pricing gained OpenAI rows (gpt-5-codex, gpt-5.3-codex, and the gpt-5.6 family), read from OpenAI's published prices on 2026-09-06.

Fixed

• The share dialog's preview never changed when you switched period — and on a real machine it never appeared at all. Two faults in the same box. The draw released its image on teardown, so changing period released the one still on screen, and a released image leaves the previous picture showing. Under that, anything the drawing threw on killed the draw silently and left "drawing…" on screen for good, which looked exactly like a slow one. The card you saved was always right; the preview was a lie about it. A draw that cannot finish now says so.

Removed

• The "Lots of turns, few files" alert. It counted files touched through Edit/Write only, so everything changed through Bash — sed -i, a heredoc, a python one-liner — was invisible to it. An agent running with permissions bypassed is told to prefer exactly those, so the rule was blind to an ordinary way of working.

On the owner's machine it had fired three times ever, on sessions that made 636, 418 and 288 Bash calls. All three were wrong, and one of them had shipped three releases while the banner said "no files touched". A rule with no true positives is not a threshold to tune, so it is gone rather than retuned. Counting files out of shell command text would be guesswork; if this ever returns it should be built on what git actually says changed.