Observed

Compression config works as documented for static models, but `ContextCompressor.threshold_percent` is not updated on runtime model switch (`/model`, fallback activation, or primary restore) — `context_length` updates while `threshold_percent` retains the previous model's value, so compression triggers at the wrong token threshold (#18617, OPEN). Earlier dossier text that described integer `threshold: 75` being silently read as "~75 tokens" is NOT the bug #18617 tracks; that integer-coercion mechanism remains unverified.

1. UPDATE 2026-07-09(v0.18.2): #18617 re-confirmed OPEN (title and body match the per-model threshold_percent sync failure). Status remains partial.
2. UPDATE 2026-07-17(v0.18.2, tag v2026.7.7.2): #18617 CLOSED 2026-07-12 by hermes-sweeper (not_planned/implemented_on_main), citing commit 76381e2a8 ("stop compaction thrash", #60989), which re-derives `threshold_percent` on `update_model()` "in both directions." Commit content confirmed via `gh api`: it touches `agent/context_compressor.py` and adds `tests/agent/test_compression_small_ctx_threshold_floor.py`, exercising exactly this recalculation — the fix's substance matches the reported bug. However, the commit (2026-07-08) is 41 commits ahead of the v0.18.2 tag (v2026.7.7.2) — merged to `main` but not shipped in any tagged release as of 2026-07-17. Falsification criterion is met on `main` but not on a released version yet; status remains partial pending a tagged release that includes it.