0d97ccccfa65764150e7976ee697bf47a9bd578e
4
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
  mayatnikovandClaude Opus 4.7
|
0d97ccccfa |
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>
|
||
|
mayatnikovandClaude Opus 4.7
|
ddc67a5031 |
v0.3.0-rc.2: manifesto / needs ladder L0-L10
Adds an explicit hierarchical needs catalogue that the reflex consults
on every tick. The bot now pursues tangible intermediate goals (food,
wood tools, shelter, stone tools, ...) instead of inheriting whatever
the curriculum thought was "next".
Ladder:
L0 alive HP>5, food>0, not in lava, not panic-near hostile
L1 food ≥6 food items in inventory (or sated + any food)
L2 tools_wood wooden_pickaxe + wooden_axe + wooden_sword
L3 shelter_basic bed placed nearby or in inventory
L4 tools_stone stone-tier triplet
L5 armor_basic any chestplate (pursue=null until craft.leather-*
lands; ladder gracefully skips)
L6 food_security ≥16 food items
L7 tools_iron iron-tier triplet (pursue=gather.stone for now)
L8 armor_iron iron chestplate (pursue=null for now)
L9 village_seed bed + chest in nearby blocks
L10 village_full never detected, falls through to curriculum
Each need has detect(snapshot) → bool and pursue(snapshot) →
{skillId, args} | null. The ladder picks the LOWEST unsatisfied
pursuable need. Needs whose pursue is null get recorded as
blockedNeeds and the walk continues — no stalling on missing skills.
Wired into curriculumReflex: manifesto takes precedence over
curriculum.plan when it has a concrete suggestion. Tests can pass
ctx.disableManifesto=true to exercise the curriculum branch
in isolation (existing reflex tests keep passing this way).
Pi self-reflection prompt now includes
"activeNeed (Maslow ladder L0-L10): L2 tools_wood → gather.logs"
so Pi advises at the right level instead of giving generic guidance.
skillId returned by pursue() is validated against the live registry
(rc.1 plumbing) — manifesto cannot accidentally dispatch a
hallucinated skill name.
Tests: 315 green (was 279 on rc.1, +36 new):
- runtime/manifesto/needs.test.js — 24 tests (per-need detect/pursue,
helper sums)
- runtime/manifesto/state.test.js — 10 tests (ladder walk, hostile
takeover at L0, armor skipping, caching)
- runtime/reflex.test.js — 2 integration tests (manifesto overrides
curriculum plan; well-fed bot pursues tools_stone)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
|
||
|
mayatnikovandClaude Opus 4.7
|
fcfa2277ba |
v0.3.0-rc.1: live skill registry + fast advisor scaffold
Roots out the v0.2.x failure mode: Pi-extracted lessons routinely named
hallucinated skill ids (relocate.surface, choose.safe.surface,
survive.shelter, gather.visible_log, …). All 47 Pi-lessons in the live DB
had applied_count=0 because normalisePreferSkill couldn't find them.
Fix:
1. runtime/skill-registry.js — single source of truth derived from
skills/index.js. Exports listSkillIds, isRegistered, and a
prompt-ready block (skillRegistryPrompt) grouped by namespace.
2. Pi prompts (coach/postmortem, coach/reflect) embed the live registry
with a "USE ONLY THESE, never invent" instruction. Lessons are
filtered at write-time too — anything not in the registry and not a
known mode name gets dropped.
3. coach/advice.js — normalisePreferSkill now returns null for unknown
ids, hardening consult() against any hallucinations that slip
through. Warn-logged for visibility.
Also lays the LLM substrate for the rest of v0.3.0:
- runtime/llm/provider.js — OpenAI-compatible chat client. Configured
via PEPA_FAST_LLM_{BASE_URL,API_KEY,MODEL,TIMEOUT_MS}. Safe no-op
unless API_KEY is set. Supports JSON-mode.
- runtime/coach/fast-advisor.js — tactical advisor tier (scaffold).
Exposes advise() that asks the fast LLM what to do RIGHT NOW when
the reflex is wedged/stuck. Rejects hallucinated skill ids using the
registry. Rate-limited 6/h, 30s cooldown. Not auto-triggered yet —
wired into reflex in rc.3 (awareness layer).
Tests: 279 green (+24 vs rc.3): 5 registry, 9 provider, 10 advisor.
See dev/v0.3.0/PLAN.md for the full iteration design (manifesto needs
ladder, event-driven awareness, skill pre-emption) and STATUS.md for
shipped/pending tracking.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
|
||
|
|
e84148d189 |
v0.2.0-rc.2: P0 hardening — Pi headless, test state isolation, advice fixes (#21)
P0 (correctness):
1. PEPA_HEADLESS=1 guard in extensions/mineflayer-bridge.ts. When `pi -p`
spawns a subprocess (banter, coach, planner, reflect, auto-patch), the
bridge no longer attempts a second MC connect — the hybrid runtime
already owns the nickname. runtime/pi-bridge.js sets the env var on
every spawn. Root cause of the "two pepa_bot's racing for the slot"
bug seen in reply-pi stderr.
2. Test state isolation in runtime/config.js. When running under the node
test runner (detected via execArgv/argv) — or when PEPA_STATE_DIR is
set — stateDir redirects to /tmp/pepa-test-state-<pid>/. log.js,
scenario-memory, world-journal, and knowledge.db all follow.
`npm test` no longer pollutes live scenarios.jsonl, world-journal.jsonl,
or daily log files. Verified empirically: post-fix run added 0 test
rows to the live scenarios file. Cleaned ~550 historical test rows
from live state in the same change.
3. defendReflex outcome reporting (runtime/reflex.js). Previously a
creeper-rule override marked the lesson succeeded=false BEFORE the
flee skill returned. Now dispatchDefendFlee accepts {lessonId} and
the onComplete fires reportAdviceOutcome with the actual flee result.
4. Mode-name → skill-id translation in runtime/coach/advice.js. Pi-coach
occasionally returns prefer_skill values that are mode names
("night_shelter", "self_preservation", "hunger"). normalisePreferSkill
maps these to SAFE_OVERRIDES entries before dispatch. Also handles
"tunnel-out", "survive_flee", "survive flee" shapes.
New behavior:
5. Self-reflection loop (runtime/coach/reflect.js). Every 30 min, the
bot asks Pi: "Are you making progress, or stuck in a loop? What
should you do differently?" Pi answers with a verdict
(progress/loop/recovering/idle/emergency), summary, next-action, and
0-N new lessons. The reflection is written to
state/<host>/reflections/<ts>.md and lessons land in the DB with
source="pi-reflect". Rate-limited to 2 calls/hour. Wired through
bot.js with the existing askPi + lastSnapshot accessor.
Tests: 246/246 green (+9 new: 6 advice mode-name + 4 reflect).
Co-authored-by: Yuriy Mayatnikov <mayatnikov@me.com>
Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
|