@nexus-cortex/cli 4.124.58 → 4.124.60

package/.env.defaults

@@ -1044,6 +1044,17 @@ CORTEX_ENDTURN_GATE=true
 #   the disclosed max. Cards now 393,216 → an unset lever sends 393,216. Use this lever to pin a LOWER cap (e.g. 65536 to reproduce the
 #   old behaviour as a control, 131072 = the provider's max-effort default). Pairs with CORTEX_HTTP_TIMEOUT_MS (below).
 #CORTEX_OUTPUT_CAP_TOKENS=
+# CORTEX_WALL_RETRY_LADDER / CORTEX_WALL_FORCE_ACTION_AT — HB-WALL-RETRY-LADDER (R241/R240, 2026-10-06, DARK). Under the 393,216 cap every
+# wall of the 10-05 TB4 cell was a degenerate repetition loop (40/40) and the retry re-looped from its first tokens at full price; DeepSeek has
+# no server-side cycle breaker in thinking mode (probed 10-06: penalties deprecated, temperature inert, stop sequences never reach
+# reasoning_content). LADDER = max_tokens for the retry after the n-th exhaustion wall of a turn ("65536,16384": wall 1 → 65,536, wall 2+ →
+# 16,384; last rung repeats; never above the request's own cap). FORCE_ACTION_AT (default 2) = from that wall on the retry carries
+# tool_choice:required, which the gateway turns into thinking OFF on DeepSeek → the model must emit a tool call (probed 10-06: works
+# mid-conversation, thinking resumes next call; costs ONE uncached prefix read). Exhaustion kinds only, never a no-tools nudge. R240 rides
+# along: the retry's usage is now counted in the session ledger. Events: wall_retry_ladder {wallIndex, cap, forceAction, iteration}.
+# [LEDGER 2026-10-06 @4.124.60] DARK — BUILT + unit/integration-tested (wallRetryLadder.test.ts, wallGuard.integration.test.ts); first arm pending.
+#CORTEX_WALL_RETRY_LADDER=
+#CORTEX_WALL_FORCE_ACTION_AT=2
 # CORTEX_HTTP_TIMEOUT_MS — HTTP timeout for the NON-STREAMING chat call (2026-10-05). The openai SDK default (600 s, 2 retries) bounds the
 # whole response body; a 384K-token thinking turn decodes 20–70 min, so the default would abort it and retry at double cost. Unset =
 # derived from the request's max_tokens at 50 tok/s with a 600-s floor (65,536 → 1,311 s; 393,216 → 7,864 s); integer ms overrides.
@@ -1279,6 +1290,9 @@ CORTEX_ENDTURN_GATE=true
 # [LEDGER 2026-10-05 mechanism] t4r trajectories: rule judge meets=false 32/36 (89%) → 14 vetoes + 11 accept-with-gap; sessions ended
 #   MID-VETO at remFrac ≥ 0.6 (cls 24,936 s). A never-satisfied gate on long budgets; harmless on 900-s MiMo only because the budget rules
 #   cut it off. Needs a fit gate + rule-veto cap before any revival. r-tb4-valA-2026-10-04.md addendum M1.
+# [LEDGER 2026-10-05 @4.124.59] INPUTS rule given an EXIT: "re-read once; if the conflict is real, use the task's authoritative source, record
+#   the discrepancy, proceed — do not keep re-reading"; judge: a recorded, justified discrepancy is not a gap. Cause: atrx-vep-crispr (t4r 10-04)
+#   looped 9 turns / 3 walls on a real 10,000-bp typo in the provided CDS file because the rule forbade the correct conclusion. Untested.
 #CORTEX_RULE_INPUTS=
 # R212 INTERP = L2 interpretation enumeration: named alternatives each computed; keep the one consistent with every explicit hint and all data.
 # [LEDGER 2026-10-02 @4.124.43] DARK — BUILT + unit-tested (taskRules.test.ts). NO bench arm yet. evidence: tb4-never-passed-2026-10-02 (foodstuff-beta-activity, glycan-ms2-elucidation, hof-topology-interpenetration, ontology-kg-querying).

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@nexus-cortex/cli",
-  "version": "4.124.58",
+  "version": "4.124.60",
   "description": "Nexus Cortex CLI - Terminal interface for multi-provider LLM orchestration",
   "main": "dist/index.js",
   "type": "module",
@@ -19,7 +19,7 @@
     "prepack": "node ../../scripts/copy-pkg-cortex-scaffold.mjs"
   },
   "dependencies": {
-    "@nexus-cortex/core": "4.124.58",
+    "@nexus-cortex/core": "4.124.60",
     "chalk": "^4.1.2",
     "cli-spinners": "^2.9.0",
     "commander": "^11.0.0",