feat: L10 Proteus support in glm.jinja

templates/glm.jinja

@@ -23,31 +23,31 @@ Operational Directives:
 
 {%- set inst_hermes -%}
 Reasoning effort: Hermes — [L1] (rapid instinct).
-Mindset: High-velocity execution. Rapid sanity check of premise, arithmetic, or syntax — then commit immediately to the shortest verified bridge to the conclusion. Conclude in a light single-digit share of output; zero when obvious.
+Mindset: High-velocity execution. Rapid sanity check of premise, arithmetic, or syntax — then commit immediately to the shortest verified bridge to the conclusion. Conclude immediately once obvious.
 {%- endset %}
 {%- set inst_apollo -%}
 Reasoning effort: Apollo — [L2] (disciplined deduction).
-Mindset: Structured, transparent logic. Isolate given constraints, trace linear implications with zero detour, and exit thinking the instant premise validation is reached (typically 10-25% of output).
+Mindset: Structured, transparent logic. Isolate given constraints, trace linear implications with zero detour, and exit thinking the instant premise validation is reached.
 {%- endset %}
 {%- set inst_artemis -%}
 Reasoning effort: Artemis — [L3] (boundary & failure audit).
-Mindset: Rigorous boundary validation. Audit null/empty states, numeric limits, concurrency hazards, off-by-one errors, and malformed inputs. Fortify against edge cases, then terminate thinking immediately (roughly 20-40% of output).
+Mindset: Rigorous boundary validation. Audit null/empty states, numeric limits, concurrency hazards, off-by-one errors, and malformed inputs. Fortify against edge cases, then terminate thinking immediately.
 {%- endset %}
 {%- set inst_athena -%}
 Reasoning effort: Athena — [L4] (strategic architecture).
-Mindset: Modular synthesis. Deconstruct requirements into structured sub-tasks with strict prerequisite dependency chains. Weigh concrete trade-offs (performance vs maintainability, latency vs throughput) and execute the optimal vector (roughly 30-50% of output).
+Mindset: Modular synthesis. Deconstruct requirements into structured sub-tasks with strict prerequisite dependency chains. Weigh concrete trade-offs (performance vs maintainability, latency vs throughput) and execute the optimal vector.
 {%- endset %}
 {%- set inst_prometheus -%}
 Reasoning effort: Prometheus — [L5] (defensive engineering).
-Mindset: Second-order systemic modeling. Anticipate upstream crashes, payload corruption, scaling cliffs, memory footprints, and lifecycle debt. Embed fault tolerance and graceful degradation directly into the design (roughly 40-60% of output).
+Mindset: Second-order systemic modeling. Anticipate upstream crashes, payload corruption, scaling cliffs, memory footprints, and lifecycle debt. Embed fault tolerance and graceful degradation directly into the design.
 {%- endset %}
 {%- set inst_solstice -%}
 Reasoning effort: Solstice — [L6] (deep systemic derivation).
-Mindset: Axiomatic derivation. Establish fundamental invariants, formulate distinct candidate architectures, stress-test each against boundary constraints, and converge on the dominant derivation with zero assumptions taken on faith (roughly 50-70% of output).
+Mindset: Axiomatic derivation. Establish fundamental invariants, formulate distinct candidate architectures, stress-test each against boundary constraints, and converge on the dominant derivation with zero assumptions taken on faith.
 {%- endset %}
 {%- set inst_hyperion -%}
-Reasoning effort: Hyperion — [L7] (first-principles synthesis; default).
-Mindset: Comprehensive first-principles synthesis. Decompose problems to atomic ground truths. Maintain continuous forward momentum across all proof surfaces. Immediate exit the instant resolution is achieved (typically 60-85% of output).
+Reasoning effort: Hyperion — [L7] (first-principles synthesis).
+Mindset: Comprehensive first-principles synthesis. Decompose problems to atomic ground truths. Maintain continuous forward momentum across all proof surfaces. Immediate exit the instant resolution is achieved.
 {%- endset %}
 {%- set inst_einstein -%}
 Reasoning effort: Einstein — [L8] (theoretical mastery).
@@ -57,6 +57,17 @@ Mindset: Invariant reduction and breakthrough synthesis. Strip peripheral comple
 Reasoning effort: Oracle — [L9] (proof-grade authority).
 Mindset: Exhaustive formal verification. Atomic proofs, invariant proofs, complete edge falsification, and 100% complete implementation without placeholders. Think as long as genuine proof requires; close immediately once established.
 {%- endset %}
+{%- set inst_proteus -%}
+Reasoning effort: Proteus — [L10] (dynamic adaptive reasoning; default).
+Mindset: Fluid cognitive adaptation. Zero predetermined depth or token quotas. Allocate only the exact reasoning required by the problem's verified constraints:
+- Low Complexity (Direct questions, lookups, syntax, simple facts): Conclude reasoning immediately and output the answer directly with zero overhead.
+- Moderate Complexity (Refactoring, multi-step algorithms, debugging, edge cases): Engage targeted validation. Audit boundary constraints, construct the solution, and exit.
+- High Complexity (Axiomatic proofs, novel systems architecture, hard mathematics): Formulate first-principles derivations. Trace every invariant until proven.
+Operating Directives:
+- Draw dynamically from whichever cognitive modes are needed: boundary falsification, structural trade-offs, defensive engineering, or foundational axioms.
+- Eliminate all self-conversational chatter ("let's see", "hmm", "wait"). Move strictly forward.
+- Terminate thinking immediately the instant the solution is definitively established and verified. Correctness is the sole boundary.
+{%- endset %}
 {%- set inst_swarm -%}
 Reasoning effort: Swarm — 20-perspective divergent swarm (specialized mode).
 Mindset: Four-phase convergent breakdown. Phase 1: Disassembly to core parameters. Phase 2: Divergent multi-angle simulation (logician, systems engineer, red-teamer, optimizer, critic). Phase 3: Cross-pollination and tension resolution. Phase 4: Convergence onto master deliverable.
@@ -96,11 +107,13 @@ Output contract: Conclude thinking and open final response with "[COUNCIL VERDIC
         {%- set store.reasoning_effort = tag_inner -%}
     {%- endif -%}
 {%- endfor -%}
-{%- set target_mode = store.reasoning_effort | default(thinking_level|default('7')) | string | lower | trim -%}
+{%- set target_mode = store.reasoning_effort | default(thinking_level|default('10')) | string | lower | trim -%}
 {%- set current_thinking = store.enable_thinking -%}
 {%- if is_explicit_regular or target_mode in ['0', 'mortal', 'disabled', 'none', 'off', 'zero', 'direct', 'chat', 'regular', 'vanilla', 'standard', 'default', 'clean'] -%}
     {%- set current_thinking = false -%}
     {%- set reasoning_instructions = '' -%}
+{%- elif target_mode in ['10', 'proteus', 'dynamic', 'auto', 'adaptive', 'fluid', 'default'] -%}
+    {%- set reasoning_instructions = inst_proteus -%}
 {%- elif target_mode in ['1', 'hermes', 'ulow', 'ultra-low', 'micro', 'minimal'] -%}
     {%- set reasoning_instructions = inst_hermes -%}
 {%- elif target_mode in ['2', 'apollo', 'low', 'compact', 'fast'] -%}
@@ -124,7 +137,7 @@ Output contract: Conclude thinking and open final response with "[COUNCIL VERDIC
 {%- elif target_mode in ['deep-swarm', 'oracle-swarm', 'council', 'amax-swarm', 'deep-research'] -%}
     {%- set reasoning_instructions = inst_deep_swarm -%}
 {%- else -%}
-    {%- set reasoning_instructions = inst_hyperion -%}
+    {%- set reasoning_instructions = inst_proteus -%}
 {%- endif -%}
 {%- if reasoning_instructions -%}<|system|>{{ reasoning_instructions }}{%- endif -%}
 {%- if tools -%}