@nexus-cortex/cli 4.124.45 → 4.124.47

package/.env.defaults

@@ -557,7 +557,19 @@ CORTEX_ALLOW_CMD_SUBSTITUTION=true
 # R37/R38 benchmarks: hybrid (true) is 20-26% faster than client-only with equal accuracy.
 ENABLE_SERVER_SIDE_TOOLS=true
 # Default API pattern for XAI models at registration time (ENABLE_SERVER_SIDE_TOOLS=true
-# overrides at request time). Options: messages | responses
+# overrides at request time). Options: messages | responses | chat
+#   chat = https://api.x.ai/v1/chat/completions (Bearer XAI_API_KEY) through the SAME chat/completions builder DeepSeek uses
+#   (append-only prefix cache, reasoning_content replay, x-grok-conv-id header). Client tools only: a chat-mode card registers
+#   no server-side tools, so ENABLE_SERVER_SIDE_TOOLS never re-routes it to /v1/responses. reasoning_effort is sent only for
+#   cards that declare an effort (grok-4.3-class; request/CORTEX_ACTION_EFFORT wins, max→high) — never grok-build-0.1 (400).
+#   Applies only to UN-PINNED xAI cards (grok-4.6 / grok-build-0.1 pin messages). Bench arms should select the pinned cards
+#   instead: `grok-4.3-chat` / `grok-build-0.1-chat`. Streaming usage needs CORTEX_STREAM_USAGE=on (same as DeepSeek); the
+#   system prompt rides CORTEX_DELIVER_SYSTEM_PROMPT (same R63 gate as every chat/completions provider).
+# [LEDGER 2026-10-03 @4.124.47] DARK — BUILT: XAI_API_MODE value `chat` + cards grok-4.3-chat / grok-build-0.1-chat (third xAI
+#   transport, additive; /v1/messages path untouched — prefix harness xai-messages rows byte-identical off/on). Unit-tested
+#   (xaiChatTransport.test.ts) + prefix rows xai-chat(grok-4.3-chat|grok-build-0.1-chat) append-only. Live 2026-10-03: grok-4.3 (effort high)
+#   + grok-build-0.1 (no effort) both finish tool_calls. NOTE: xAI completion_tokens EXCLUDES reasoning_tokens (DeepSeek's include
+#   them) — canonical outputTokens under-counts billed output here; use reasoningTokens / costUsd (cost_in_usd_ticks). Never benched.
 XAI_API_MODE=messages
 # OpenAI request surface. Options: chat/completions | responses (opt into hosted tools)
 OPENAI_API_MODE=
@@ -1936,6 +1948,83 @@ MENTORSHIP_ACTIVE_DISCOVERY=false
 # accepted stream_options (200; usage on the final chunk WITH finish_reason, not a separate choices:[] chunk). The prefix
 # harness cannot reach the streaming builders.
 #CORTEX_STREAM_USAGE=
+# CORTEX_OPENAI_TOOLS_REASONING — HB-OPENAI-TOOLS-REASONING (DARK, 2026-10-02; stack-portability N1). OpenAI reasoning cards never declare
+# reasoning.toggleable, so the gateway withheld the request/action effort (CORTEX_ACTION_EFFORT was a NOOP on every OpenAI path).
+# on: the gateway forwards the effort for OpenAI reasoning cards → Responses `reasoning.effort` (+ summary auto). Chat/completions is
+# UNCHANGED: R19b's drop stays — live probe 2026-10-03, gpt-5.4-nano + gpt-5.6-luna: chat reasoning_effort + function tools = 400
+# "use /v1/responses"; Responses + tools + reasoning.effort = 200 on both. So an OpenAI effort arm must use OPENAI_API_MODE=responses.
+# Values pass through verbatim. DeepSeek and every non-OpenAI request byte-identical. off: unchanged.
+# [LEDGER 2026-10-03 @4.124.46] DARK — BUILT; chat half REMOVED after the live probe (would have 400ed every OpenAI tool call);
+# Responses half live-verified (accepts effort with tools).
+#CORTEX_OPENAI_TOOLS_REASONING=
+# CORTEX_ANTHROPIC_EFFORT — HB-ANTHROPIC-EFFORT (DARK, 2026-10-02; stack-portability N1). Adaptive-thinking families (claude-4.7,
+# claude-4.8, claude-5, claude-fable-5) got thinking {type:adaptive} only — the action effort never reached the wire; budget
+# families mapped effort `max` to the 10000 fallback (not in the budget map). on: adaptive families send
+# `output_config: {effort}` = low|medium|high|xhigh|max (GA field, no beta header; none/unset → omitted = provider default);
+# budget families map `max` to budget_tokens = min(120000, max_tokens - 1024) (must stay < max_tokens and >= 1024; below
+# that the legacy 10000 stays). low/medium/high budgets unchanged. Send + stream paths. xAI Messages untouched. An effort
+# change mid-session invalidates Anthropic's messages cache (same as the existing budget ladder). off: unchanged.
+# [LEDGER 2026-10-02 @4.124.46] DARK — BUILT + unit-tested (adaptive + budget, send + stream, off = byte-identical) + prefix
+# harness (send a, ACTION_EFFORT=max: claude-sonnet-5 output_config none → {effort:max} on 6/7 main requests; sonnet-4.6 and
+# haiku-4.5 budget_tokens 10000 → 62976 (max_tokens 64000) on 6/7; pair verdicts unchanged). No live Claude probe.
+#CORTEX_ANTHROPIC_EFFORT=
+# CORTEX_GEMINI_TOOLS_THINKING — HB-GEMINI-TOOLS-THINKING (DARK, 2026-10-02; stack-portability N1). The @google/genai builders
+# skipped thinkingConfig whenever tools were attached ("cannot coexist", a Gemini 2.0-era observation) and the REST builder only
+# sent a card-level thinkingLevel, so CORTEX_ACTION_EFFORT never reached Gemini. on: the gateway maps the request/action effort
+# to generationConfig.thinkingConfig on REST + SDK, with or without tools — Gemini 3.x thinkingLevel (low|medium|high;
+# xhigh/max → high), 2.5 and older thinkingBudget (low 2048, medium 8192, high 16384, xhigh/max 24576); the request effort
+# replaces a card-level thinkingLevel. SDK + tools with no effort → thinkingBudget 10000 (the no-tools default). Streaming SDK
+# adds includeThoughts; non-streaming does not (the message-level adapter would render thought parts as visible text). Visible
+# reasoning (wall levers) is NOT enabled by this flag. off: unchanged.
+# [LEDGER 2026-10-02 @4.124.46] DARK — BUILT + unit-tested (REST 3.x level + 2.5 budget, SDK send + stream, off =
+# byte-identical) + prefix harness (send a, ACTION_EFFORT=max: gemini-3.7-flash REST thinkingLevel "high", gemini-2.5-flash
+# REST + SDK thinkingBudget 24576 on 6/7 main requests; pair verdicts unchanged). No live Gemini probe of thinking + tools.
+#CORTEX_GEMINI_TOOLS_THINKING=
+#
+# CORTEX_ACTION_EFFORT_FIRST — HB-ACTION-EFFORT-FIRST (DARK, 2026-10-03; found by the effort-gap build). CORTEX_ACTION_EFFORT is filled
+# into the request params AFTER request 0 of each user turn is built, so request 0 ran at the model card's default effort and requests 1+
+# at the action effort. DeepSeek: harmless (card 'medium' = 'high' server-side; the anchor-lift bust is on the same boundary). Other
+# providers: request 0 runs at the wrong effort and the effort flip costs one extra cache break per user turn (Anthropic output_config,
+# Gemini thinkingConfig, OpenAI reasoning_effort are all prefix-relevant). on = request 0 also gets CORTEX_ACTION_EFFORT (a ramp or an
+# explicit request param still wins). off = unchanged (byte-identical).
+# [LEDGER 2026-10-03 @4.124.46] DARK — BUILT (unit tests; turn it on in every non-DeepSeek cross-provider arm).
+#CORTEX_ACTION_EFFORT_FIRST=
+#
+# CORTEX_GEMINI_EXPLICIT_CACHE — HB-GEMINI-EXPLICIT-CACHE (DARK, 2026-10-03; stack-portability §"GEMINI EXPLICIT cachedContents
+# FEASIBILITY"). Gemini implicit caching is partial on gemini-3-flash-preview (fixed ~12K head only) and NONE on gemini-3.8-flash;
+# explicit cachedContents cached 13,405/13,414 on both (live 2026-10-03). on: the Gemini builders (REST send + stream, @google/genai
+# send + stream) snapshot (model, system, tools) + history up to (not including) the newest turn as a cachedContents once it reaches
+# the model minimum (3.x 4,096 / 2.5 2,048 tokens, ai.google.dev/gemini-api/docs/caching), then send cachedContent + only the turns
+# after the snapshot with NO system/tools/toolConfig (Gemini 400s that pair). Re-snapshot when the uncached tail passes
+# CORTEX_GEMINI_CACHE_RESNAPSHOT_TOKENS, when the history before the snapshot is rewritten (hash check), or near expiry; a
+# system/tools change (anchor lift) = new snapshot. Superseded snapshots are DELETEd; orchestrator cleanup DELETEs the rest. A forced
+# tool_choice request bypasses the cache. Any 400/403/404 on a cached request = the full request once + slot dropped; a failed
+# create = full request + backoff. Logs `[gemini-cache]` lines (CORTEX_GEMINI_CACHE_QUIET=true silences). Cost (pricing page,
+# 2026-10-03): creation = the prefix at the normal input rate once; cached reads 3 Flash $0.05/M (input $0.50), 3.8 Flash $0.075/M
+# (input $0.75, through 2026-12-31); storage 3 Flash $1.00, 3.8 Flash $0.50 per M tokens per hour. off: unchanged (byte-identical).
+# [LEDGER 2026-10-03 @4.124.47] DARK — BUILT (16 unit tests on a fake Gemini API; off = byte-identical send + stream) + one live
+# 5-turn tool loop on gemini-3-flash-preview + gemini-3.8-flash through the built dist (see the 4.124.47 changeset). No bench arm yet.
+#CORTEX_GEMINI_EXPLICIT_CACHE=
+#
+# CORTEX_GEMINI_CACHE_TTL_S — HB-GEMINI-EXPLICIT-CACHE: snapshot TTL in seconds (min 60). A snapshot used in the second half of its
+# life is PATCHed back to the full TTL. Storage is billed per token-hour, so the TTL bounds the cost of an orphaned snapshot.
+# [LEDGER 2026-10-03 @4.124.47] DARK — BUILT (inert unless CORTEX_GEMINI_EXPLICIT_CACHE=on). Default 300.
+#CORTEX_GEMINI_CACHE_TTL_S=300
+#
+# CORTEX_GEMINI_CACHE_RESNAPSHOT_TOKENS — HB-GEMINI-EXPLICIT-CACHE: re-snapshot a longer prefix once the uncached tail (history after
+# the snapshot, newest turn excluded; chars/4 estimate) exceeds this many tokens (min 1024). Lower = more creations (each billed at
+# the full input rate), higher = more uncached input per request.
+# [LEDGER 2026-10-03 @4.124.47] DARK — BUILT (inert unless CORTEX_GEMINI_EXPLICIT_CACHE=on). Default 16000.
+#CORTEX_GEMINI_CACHE_RESNAPSHOT_TOKENS=16000
+#
+# CORTEX_GEMINI_CACHE_MIN_TOKENS — HB-GEMINI-EXPLICIT-CACHE: override the per-model minimum cacheable size (default 4096 for 3.x
+# incl. gemini-3-flash-preview, which the docs table does not list; 2048 for 2.5).
+# [LEDGER 2026-10-03 @4.124.47] DARK — BUILT (inert unless CORTEX_GEMINI_EXPLICIT_CACHE=on). Default: per model.
+#CORTEX_GEMINI_CACHE_MIN_TOKENS=
+#
+# CORTEX_GEMINI_CACHE_QUIET — HB-GEMINI-EXPLICIT-CACHE: true silences the `[gemini-cache]` stderr lines.
+# [LEDGER 2026-10-03 @4.124.47] DARK — BUILT (inert unless CORTEX_GEMINI_EXPLICIT_CACHE=on). Default false.
+#CORTEX_GEMINI_CACHE_QUIET=
 
 # Prompt-mass composition (BASH_PLUS_SPEC P6): full (default) | minimal (only
 # the core system prompt; guides/CLAUDE.md/memory dropped) | defer (minimal on

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@nexus-cortex/cli",
-  "version": "4.124.45",
+  "version": "4.124.47",
   "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": "4.124.45",
+    "@nexus-cortex/core": "4.124.47",
     "chalk": "^4.1.2",
     "cli-spinners": "^2.9.0",
     "commander": "^11.0.0",