refactor: rename teacher/student to optimizer/target, remove best skills, fix slow update

- Rename teacher -> optimizer, student -> target across all code, configs, docs, prompts
- CLI: --teacher_model -> --optimizer_model, --student_model -> --target_model
- Remove best_skill files, keep only initial skills
- Fix slow update gate (force write into skill)
- Fix SLOW_UPDATE marker stripping
- Remove deep_reflect and meta_reflect mechanisms
- Update .env.example with export prefix and azure_cli docs
- Add endpoint empty validation in azure_openai.py

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

skillopt/optimizer/meta_skill.py

@@ -1,28 +1,28 @@
-"""Teacher-side meta skill memory for cross-epoch optimization guidance.
+"""Optimizer-side meta skill memory for cross-epoch optimization guidance.
 
-This module maintains a compact teacher-facing memory distilled from
+This module maintains a compact optimizer-facing memory distilled from
 adjacent-epoch skill comparisons. Unlike ``slow_update``, it does not
-modify the student skill document. Instead, it produces guidance meant to
-improve future teacher behavior when proposing, merging, and ranking edits.
+modify the target skill document. Instead, it produces guidance meant to
+improve future optimizer behavior when proposing, merging, and ranking edits.
 """
 from __future__ import annotations
 
 import traceback
 
-from skillopt.model import chat_teacher
+from skillopt.model import chat_optimizer
 from skillopt.optimizer.slow_update import format_comparison_text
 from skillopt.prompts import load_prompt
 from skillopt.utils import extract_json
 
 
 def format_meta_skill_context(meta_skill_content: str) -> str:
-    """Render teacher memory into a prompt-ready context block."""
+    """Render optimizer memory into a prompt-ready context block."""
     content = (meta_skill_content or "").strip()
     if not content:
         return ""
     return (
-        "## Teacher Meta Skill\n"
-        "This is teacher-side memory distilled from prior epoch transitions in "
+        "## Optimizer Meta Skill\n"
+        "This is optimizer-side memory distilled from prior epoch transitions in "
         "this environment. Use it to improve how you propose, merge, and rank "
         "skill edits. Prefer it when the current evidence is ambiguous, but do "
         "not force it if the current trajectories clearly contradict it.\n\n"
@@ -38,7 +38,7 @@ def run_meta_skill(
     prev_meta_skill_content: str = "",
     system_prompt: str | None = None,
 ) -> dict | None:
-    """Produce updated teacher-side meta skill from adjacent epochs."""
+    """Produce updated optimizer-side meta skill from adjacent epochs."""
     actual_system = system_prompt if system_prompt is not None else load_prompt("meta_skill")
 
     prev_skill_display = prev_skill
@@ -52,15 +52,15 @@ def run_meta_skill(
     prev_meta_section = (
         prev_meta_skill_content.strip()
         if prev_meta_skill_content and prev_meta_skill_content.strip()
-        else "(No previous teacher meta skill — this is the first update.)"
+        else "(No previous optimizer meta skill — this is the first update.)"
     )
 
     comparison_text = format_comparison_text(comparison_pairs)
     user = (
         f"## Previous Epoch Last-Step Skill\n{prev_skill_display}\n\n"
         f"## Current Epoch Last-Step Skill\n{curr_skill_display}\n\n"
-        f"## Previous Teacher Meta Skill\n"
-        f"The following teacher memory was available during the current epoch. "
+        f"## Previous Optimizer Meta Skill\n"
+        f"The following optimizer memory was available during the current epoch. "
         f"Reflect on whether it improved or harmed the quality of edits.\n\n"
         f"{prev_meta_section}\n\n"
         f"## Longitudinal Comparison (same tasks, two last-step skills)\n"
@@ -68,7 +68,7 @@ def run_meta_skill(
     )
 
     try:
-        response, _ = chat_teacher(
+        response, _ = chat_optimizer(
             system=actual_system,
             user=user,
             max_completion_tokens=3072,