Harness coverage
TokenShift supports five coding agent harnesses today — Claude Code, Codex, GitHub Copilot CLI, Cursor, and opencode.
Output compression, telemetry, and session-history backfill work on all five — on macOS, Linux and Windows alike. Three things differ.
What differs
Section titled “What differs”| Harness | Full-output recovery | Token usage reporting | Hook config |
|---|---|---|---|
| Claude Code | Yes | Per turn | Shared with the harness |
| Codex | No * | Per turn | Shared with the harness |
| Copilot CLI | Yes | At session end | Own file |
| Cursor | No * | Per turn | Shared with the harness |
| opencode | Yes | Per turn | Own file |
* Codex and Cursor sandbox the proxy, so there is no local cache to replay from. Compression and telemetry are unaffected — see Recovery cache scope.
What the three columns mean
Section titled “What the three columns mean”- Full-output recovery — whether
tokenshift recovercan pull back the original, uncompressed output of a call TokenShift compressed. - Token usage reporting — when the model’s own token counts reach your telemetry.
- Hook config — whether TokenShift keeps its hook wiring in its own file, or shares the harness’s. Shared is fine in normal use; worth knowing if another tool edits it too.
Every row above is a property of the harness, not of the operating system. Where a difference between operating systems appears, it will be called out here.
- How it works — what compression, telemetry, and the intent classifier each do.
- Data contract — hook-event coverage, field by field.