feat(v0.3.0): paradigm shift — TimeWeb-only LLM + persistent advisor trail + improvement queue

This is the rc.4 batch the user requested:

  1. Emergency triggers (low HP + close hostile, lava-under-foot)
     bypass the long cooldown so the LLM is consulted BEFORE the bot
     dies, not after.
  2. Active manifesto need is now included in the advisor user prompt
     — the LLM picks suggestions that satisfy the bot's current
     concrete need (L2 tools_wood → "gather logs nearby" not
     "explore further").
  3. Every advisor recommendation is persisted to SQLite
     (advisor_recommendations table) with full token usage. The
     reflex marks 'applied=1' when it dispatches and updates
     outcome_ok/code when the dispatch completes. Ground truth for
     "is the LLM actually helping" lives in the DB, not in logs.
  4. Pi CLI is OUT of every background loop. coach/postmortem and
     coach/reflect now go through the same TimeWeb endpoint
     fast-advisor uses, via the shared coach/llm-call.js helper.
     Pi is reserved for manual operator commands.
  5. The LLM (postmortem, reflect, advisor) can flag "structural
     gaps" — missing skills/features the operator should implement.
     These land in the new improvement_requests table. Dedup by
     title bumps `votes` instead of inserting duplicates so the
     queue doesn't bloat. Operator views via
     `node scripts/list-improvements.js`.
  6. A deterministic trigger-tuner runs hourly: reads 24h of
     recommendation stats, flags triggers whose success rate is
     below 25% (sample ≥ 5) or whose prompts are expensive (>1000
     input tokens) with mediocre payoff. Improvements get
     source="tuner", category="tuning". No LLM call.

New files:
  runtime/coach/llm-call.js        — askAnalytical() helper
  runtime/coach/trigger-tuner.js   — stats → improvements
  runtime/coach/trigger-tuner.test.js
  scripts/list-improvements.js     — operator CLI

Schema additions:
  advisor_recommendations: id, ts, trigger_reason, planned_skill,
    recommended_skill, action, rationale, active_need, tokens_in,
    tokens_out, latency_ms, applied, outcome_ok, outcome_code, outcome_at
  improvement_requests: id, ts, source, category, title, description,
    context, priority, status, duplicate_of, votes, implemented_at, notes

Renamed env-var consumers:
  Pi-coach drainOnce({ askPi })   → drainOnce({ askAnalyticalFn? })
  Pi-reflect runOnce({ askPi })   → runOnce({ askAnalyticalFn? })
  bot.js attachCoach/attachReflect no longer pass askPi
  attachTuner() added to bot.js spawn handler
  lessons.source 'pi-coach'   → 'timeweb-coach'
  lessons.source 'pi-reflect' → 'timeweb-reflect'

Token cost measured live:
  ~705 input + 45 output = ~750 total per advisor call
  worst case @ 6 calls/hour rate cap = ~108K tokens/day
  OpenAI gpt-5-mini reference price: ~$0.60/month

Operator usage:
  node scripts/list-improvements.js                # open queue
  node scripts/list-improvements.js --stats        # advisor performance
  node scripts/list-improvements.js --done 17 "shipped in 0.3.1"
  node scripts/list-improvements.js --reject 18 "duplicate"

Tests: 360 green (was 332, +28 new).

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

runtime/reflex.js

@@ -27,6 +27,7 @@ import {
 import { runSkill, getSkill } from "./skills/index.js";
 import { consult as consultAdvice, reportOutcome as reportAdviceOutcome } from "./coach/advice.js";
 import { tickAdvisor, consumeFreshRecommendation } from "./coach/advisor-trigger.js";
+import { markRecommendationApplied, markRecommendationOutcome } from "./knowledge/index.js";
 import { pickActiveNeed } from "./manifesto/state.js";
 import { situationHash } from "./scenario-memory.js";
 import { tickModes } from "./modes.js";
@@ -537,12 +538,15 @@ function curriculumReflex(ctx) {
 	// v0.3.0 fast-advisor: if a fresh recommendation is sitting on ctx
 	// (the result of a previous tick's async advise() call), use it.
 	// This is the closing of the awareness → LLM → action loop.
+	let appliedRecommendationId = null;
 	if (!ctx.disableAdvisor) {
 		const rec = consumeFreshRecommendation(ctx);
 		if (rec && rec.skillId) {
 			info(REFLEX_LOG, `advisor override: ${skillId} → ${rec.skillId} (${rec.triggerReason}, ${rec.rationale?.slice(0, 60)})`);
 			skillId = rec.skillId;
 			skillSource = `advisor:${rec.triggerReason}`;
+			appliedRecommendationId = rec.id ?? null;
+			if (appliedRecommendationId) markRecommendationApplied(appliedRecommendationId);
 		}
 		// Always fire-and-forget another advise() if triggers fire — the
 		// result lands on a future tick. tickAdvisor handles its own
@@ -603,6 +607,12 @@ function curriculumReflex(ctx) {
 		onComplete: (res) => {
 			ctx.skillBackoff = ctx.skillBackoff ?? {};
 			if (advice.lessonId) reportAdviceOutcome({ lessonId: advice.lessonId, succeeded: !!res?.ok });
+			if (appliedRecommendationId) {
+				markRecommendationOutcome(appliedRecommendationId, {
+					ok: !!res?.ok,
+					code: res?.code ?? null,
+				});
+			}
 			if (res?.recovery?.hint === "wander") {
 				// Same fix the old autonomous reflex applied for "no reachable
 				// log" — switch to exploration for a minute.