Real test revealed gpt-4o-mini was still:
- Hallucinating suite names ("Aluba" doesn't exist — we only have
Alexa, Stilo, Hidromassagem)
- Extracting cadastral data as memory ("Rodrigo has a CPF", "Name is X")
despite the per-type NÃO examples
Added two sections at the top of the prompt:
1. Business canonical data — explicit whitelist of suite names (Alexa,
Stilo, Hidromassagem) and stay types. Anything else = discard, NO auto-
normalization. LLM must not guess.
2. Cadastral data absolute rule — explicit list of fields that are
profile data, not memory: name, CPF/RG/passport, email/phone/address,
birth date. Plus 5 concrete ❌ examples of what was being wrongly
extracted in the wild.
Existing 9 specs still pass (stub at call_llm; prompt change is
semantic, not structural).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Real-world test revealed the LLM extractor (gpt-4o-mini) was using type
labels too loosely: a customer's QUESTION about parking ("tem
estacionamento?") was classified as 'reclamacao'. Similarly cortesia
generica ("obrigado") was becoming 'feedback_positivo', and transactional
events (CPF informed, reservation made) were becoming memories when they
should be ignored.
Rewrote build_prompt with:
- Per-type strict definition (what it IS)
- YES/NO examples for each of the 9 types, with the most common pitfalls
explicitly shown as NO
- 7 absolute rules, including: questions are never complaints, generic
courtesy is never feedback, agent actions are never customer memory,
transactional events are not long-term facts
- Confidence threshold guidance (>=0.9 only if totally explicit, 0.7-0.89
for strong inference, <0.7 discard)
- "If in doubt, discard — quality > quantity. Most transactional
conversations should return empty facts list"
Existing 9 specs still pass (stub call_llm, so prompt changes don't
affect unit test assertions).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Real-world observation: OpenAI embedding API takes 200-400ms typical,
plus pgvector query overhead, the 500ms budget was being exceeded
frequently, silently dropping memory recall. Agent typing delay is
already 2-15s humanized, so a 2s recall budget is well within UX
tolerance and gives ~4-5x margin over typical embedding latency.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Documents the Rails console procedure to toggle
captain_contact_memory_extraction_enabled and
captain_contact_memory_recall_enabled on Account#custom_attributes,
including rollout phasing (extraction-first, then recall), rollback,
bulk enablement, and post-activation verification queries.
The UI toggles in Captain Settings are deferred: the existing
FeatureToggle component is coupled to the captain_features hash and
cannot be reused for custom_attributes-backed flags without a new
component and a new account-update store action. Scope and
implementation notes for that follow-up are included at the end of the
document.
Task 5.4 of Captain Semantic Memory epic (Phase 5).
Spec do Epico A - adiciona Camada 3 (memoria semantica episodica do contato)
ao Captain AI, mantendo as 3 camadas existentes inalteradas.
Decisoes fechadas no brainstorming:
- Extracao ao resolver conversa OU silencio > 30min (100% automatico)
- Validacao: evidence obrigatoria, confidence >= 0.5 (alternativas B/C/D
documentadas como fallback)
- Scope global no recall, atribuicao por source_unit_id pra relatorios
- 9 tipos iniciais, limite 5 fatos/conversa, 50 ativos/contato
- TTL por tipo + supersedencia automatica por contradicao
- LGPD soft-30d -> hard-delete via cron
- 2 feature flags independentes, default OFF
- Epico B (LangGraph/inteligencia) sera spec separado pos-producao
Custo estimado: ~R$ 47/mes no grupo todo.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Problema observado: Daniela chamou generate_pix com arguments vazios apos
cliente informar "27/4". Tool retornou missing_fields=[check_in, amount] e
LLM caiu no fallback silenciosamente.
Correcoes:
- DDMMYYYY_REGEX agora aceita "DD/MM" sem ano (assume ano corrente, empurra
pro proximo ano se a data ja passou)
- parse_date_without_year com fallback explicito
- Instruction da scenario Daniela_Reservas (DB, scenario_id=2) atualizada
para listar todos os 4 parametros obrigatorios de generate_pix e
distinguir requires_input (erro do LLM) de success=false (erro tecnico)
Backup da instruction anterior: /tmp/daniela_instruction_backup_20260418.txt
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Inboxes without portal_id were crashing with NoMethodError on save,
blocking landing host creation via UI for any inbox without a portal.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replaces stub Settings.vue with full implementation: anti-spam guard
form (quiet hours, interval, pause-on-reply, opt-out label) and a
collapsible ConciergeUnitCard per unit (inbox selector, persona name,
knowledge base, key-value variables). Adds CONCIERGE_CONFIGURED /
CONCIERGE_NOT_CONFIGURED i18n keys to en + pt_BR.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Implements Task 13 — replaces the stub History.vue with a real paginated
table filtered by status, and adds DeliveryPreviewModal to show rendered_body.
Also extends i18n keys (TOTAL, PAGINATION, MODAL labels) in en + pt_BR.
Adds CAPTAIN_LIFECYCLE block (en + pt_BR) to captain.json with full
key set for Rules, Wizard, Settings, History and sidebar entry.
Also stages pre-existing uncommitted additions to captain.json from
prior work (KPI, PILLS, QUICK_DATE, CARD, ACTIONS extras for
CAPTAIN_RESERVATIONS) — those were already in the working tree and
belong to the same feature branch.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
