Review — cannaoil

← overview

ventana últimos 7d vs 7d previos · tz America/Argentina/Buenos_Aires · último cambio Ninjo: 2026-06-05 20:10 (deploy config, hace 56d) · 3 untracked
2
Nuevas (7d)
prior 4
0
Calls booked
— de link
0
Link sent
0.0% de activas
$0
Revenue variable
3 msgs

Resumen — período vs período

Métricathis 7dprior 7dΔ
Activas38-62.5%
Nuevas24-50.0%
Link sent01-100.0%
Calls booked00
Booked/Link %0.0%
Booked/Activas %0.0%0.0%+0.0pp
Booked/Nuevas %0.0%0.0%+0.0pp
Revenue variable$0.00$0.00

Funnel — semana actual vs anterior

PeríodoActivasNuevasLink sentCalls bookedBooked/LinkBooked/ActivasBooked/Nuevas
Semana actual (7d)32000.0%0.0%
Semana anterior (7d)84100.0%0.0%0.0%
Ver desglose por día
DíaActivasNuevasLink sentCalls bookedBooked/LinkBooked/ActivasBooked/Nuevas
07-300000
07-290000
07-280000
07-2722000.0%0.0%
07-260000
07-250000
07-2410000.0%
07-2321000.0%0.0%
07-2210100.0%0.0%
07-2121000.0%0.0%
07-2022000.0%0.0%
07-190000
07-180000
07-1710000.0%
run 2026-07-24T14:58:26+00:00 · fuente: live DB state (daily_audit pipeline)
7/10
BP semánticos PASS
top-10 best practices
10/16
Determinísticos PASS
3 FAIL · 3 WARN
C0 · H0 · M2
Fails por severidad
critical / high / medium
Drift
DB vs repo local

BP semánticos (top-10 best practices)

BP-1
BP-2
BP-3
BP-4
BP-5
BP-6
BP-7
BP-8
BP-9
BP-10
BPSevEstadoDetalle
BP-1HIGHPASSTemplate compliance
meta.md declares template `universal_dm`, but system.md is structured as a `principles` template (Identity, Voice, Flow, Calibration, Principles, Tools, NO_RESPONSE). The 16 sections required by `universal_dm` (Entry Types, Path Routing, 7-phase flow detail, Qualification Scorecard, Show-Up Protection, Follow-Up Logic, etc.) are absent or severely abbreviated.
evidence (3)
  • {'file': 'meta.md', 'quote': '**Template**: universal_dm'}
  • {'file': 'system.md', 'quote': '## Identidad\n## Voz\n## Flujo\n## Gate A\n## Keywords\n## Calibración\n## NO_RESPONSE\n## Principios'}
  • {'file': 'universal_dm template spec', 'quote': 'Sections (in order): 1. Identity, 2. Tone and Style, 3. Conversation Length — Hard Cap, 4. Hard Invariants, 5. Mandatory Qualification Gates, 6. Disqualification Rules, 7. Entry Types and Routing, 8. Path Routing, 9. Universal DM Flow — 7 Phases, 10. Path-Specific Logic, 11. Non-Fit Fallback, 12. Qualification Scorecard, 13. Show-Up Protection, 14. Follow-Up Logic, 15. What to Never Do, 16. Links Reference Table'}
BP-2MEDIUMFAILExample coverage and cap (7 base scenarios + conditional, ≤20 total)
examples.md contains only 1 worked example (price inquiry shortcut). The rubric requires ≥6 of 7 base scenarios (happy path, resource/link delivery, objection handling, graceful disqualification, emotional/sensitive lead, re-engagement, off-topic question) plus conditional scenarios (keyword trigger, price objection). Only 1 example is far below the minimum coverage threshold.
evidence (2)
  • {'file': 'examples.md', 'quote': '## Ejemplo 1: Lead pregunta precio directo\n**Lead**: ¿cuánto vale?\n...\n**Agente**: ¿De dónde nos escribís, CABA o zona norte?'}
  • {'file': 'rubric BP-2', 'quote': 'Pass criterion: ≥6 of 7 base scenarios represented, every applicable conditional scenario has at least one example, AND total examples ≤20.'}
BP-3HIGHPASSExamples align with the flow
The single example in examples.md (Ejemplo 1) ends mid-flow without a terminal state. The agent's last turn asks for zone ('¿De dónde nos escribís, CABA o zona norte?'), which is a coordination step (Paso 5), but the conversation does not reach a terminal state (payment confirmed, NO_RESPONSE triggered, or disqualification). The example is incomplete.
evidence (2)
  • {'file': 'examples.md', 'quote': '**Agente**: Perfecto. Para reservar el precio de hoy te paso el alias y coordinamos la entrega. ¿De dónde nos escribís, CABA o zona norte?'}
  • {'file': 'rubric BP-3', 'quote': 'Pass criterion: Each example walks the flow without skipping required steps and reaches a terminal state.'}
BP-4CRITICALPASSNo internal contradictions in system.md
system.md contains no self-contradictory rules under the same precondition. Paso 0 (connection, 2-3 questions) and Paso 1 (minorista/mayorista classification) are sequential, not simultaneous. Gate A (cases of serious illness) and the standard flow (Paso 2 diagnosis) apply to different lead types. No shared precondition produces conflicting outputs.
BP-5CRITICALPASSNo cross-file contradictions
All SDK files agree on behavior, parameters, and data. program.json declares `booking_url: null` (no booking link, direct chat sale), which aligns with system.md's Paso 5 (transfer to alias `purasanacion24`). Pricing in program.json ($75.000 combos, $36.000 aceite) matches knowledge_base.md and examples.md. No file demonstrates behavior that contradicts another under the same precondition.
BP-6MEDIUMN/AVoice consistency across examples
Only one example exists in examples.md, so voice consistency cannot be assessed across multiple examples. This rule requires at least two examples to compare.
BP-7CRITICALPASSUnguarded actions (PAT-004)
The primary action (passing the payment alias `purasanacion24`) is guarded by the explicit precondition in Paso 5: 'confirmá dirección y zona ANTES de dar total.' The action is not unguarded. Gate A (consulta médica $60.000) is also guarded: triggered only when the lead mentions serious illness (cáncer, SIDA, autoinmunes, VIH, lupus, esclerosis, Parkinson) or multiple medications.
BP-8HIGHPASSNo stale version language in examples
The single example in examples.md uses current language from system.md (Paso 0–5 flow, alias `purasanacion24`, zone-based pricing). No stale or outdated phrasing detected.
BP-9MEDIUMFAILNo duplicated instructions
The instruction 'confirmá dirección y zona ANTES de dar total' appears in both system.md (Paso 5) and knowledge_base.md (Envíos section: 'Flujo obligatorio: pedir dirección ANTES de dar total'). This is a duplicated behavioral instruction that should have a single canonical home with a cross-reference in the other.
evidence (2)
  • {'file': 'system.md', 'quote': '**5. Coordinación de entrega**: confirmá dirección y zona ANTES de dar total.'}
  • {'file': 'knowledge_base.md', 'quote': 'Flujo obligatorio: **pedir dirección ANTES de dar total** (sin zona no se calcula envío).'}
BP-10MEDIUMPASSRule-vs-principle balance
Behavioral guidance (tone, flow, conversational style) is expressed via Voz section and Principios section, supported by the single example. Hard rules (pricing, envío zones, Gate A triggers) remain as rules in Calibración and Knowledge Base. The balance is appropriate.

Checks determinísticos

6a
6c
6d
6e
6g
6h
6i
6j
6k
6l
6m
6n
6q
6r
6s
6t
CheckSecEstadoNotas
6a1.8PASSNO_RESPONSE present (limits not checked, --no-db)
6c3.xWARN163 lines
6d1.2FAIL1 examples (target 10, min 7)
6e1.3FAIL11 principles, cognitive overload
6gPAT-006WARNnon-STEP-0 keyword handling
6hAnti-patPASSno 'burbuja'
6iAnti-patPASSno HTML comments
6jAnti-patFAIL9 em-dashes across 3 file(s); first: system.md:93
6kPAT-006PASSno tool syntax leak
6lResourcesPASSno placeholder URLs
6mHygienePASSno audit metadata in prompt files
6nPAT-008PASSURL forms consistent across files
6qPAT-001PASSno instruction duplication into examples.md, keywords.json, program.json, resources.json
6rKeyword-driftPASSall 14 keyword(s) consistent across keywords.json + system.md (no DB)
6sHygienePASSno expired dates in live SDK
6tCoverageWARN11/11 principle(s) without example: #1 Recomendá siempre del catálogo, #2 Escuchá antes de vender, #3 Mencioná el precio con, #4 Cerrá en el chat, #5 Si pregunta algo que
canal · sync 2026-07-21
95
Digests
0 en últimos 14d
0
Cambios pedidos / pendientes
del checklist
96
Threads
en el canal
2026-07-21
Última sync
del digest

Cosas nuevas (digests recientes)

FechaTemaTags
2026-07-16combo messaging naturalness iterationprompt-iteration, sales-copy, combo-products, ux-feedback
2026-07-13weekly funnel report jul 6 12 critical volume dropmetrics, alert, funnel-performance, purasanacion.arg
2026-07-13weekly funnel report jul06 jul12 critical volume dropautomated-report, metrics, alert
2026-07-06weekly funnel report jun29 jul05 volume surgeanalytics, performance, conversion, content-strategy
2026-07-06weekly funnel report jun29 jul05 189pct growthmetrics, performance, keyword-analysis, content-strategy
2026-07-03urgency sales copy 40 percent combo july 3 4sales-strategy, copywriting, promotional-campaign, urgency-pricing
2026-07-01sales closing strategy implementationsales-strategy, closing, follow-up, urgency
2026-07-01sales closing strategies implementationsales, strategy, agent-prompt, closing, urgency, follow-up
2026-06-29weekly funnel report jun22 28metrics, conversion, content-strategy
2026-06-29weekly funnel report jun22 jun28reporting, metrics, conversion, content-strategy
2026-06-26brand identity and credibility messagingdeployment, brand, identity, prompt-update
2026-06-26location and shipping info standardizationdeployment, logistics, faq, prompt-update
2026-06-26brand credibility messaging standardizationprompt-update, branding, all-agents, deployed
2026-06-26age question removal legal ong messagingprompt-update, legal-compliance, pending
2026-06-26location shipping faq standardizationprompt-update, logistics, all-agents, deployed
2026-06-25downsell strategy post combofeature-request, conversion-optimization, prompt-update
2026-06-25urgency pricing and followup logicdeployment, pricing, urgency, prompt-update
2026-06-25downsell post combo logic deploymentdeployment, pricing, downsell, prompt-update
2026-06-24whatsapp blocking contact lossbug, critical, whatsapp, contact-loss
2026-06-22weekly funnel report 56 percent growthmetrics, conversion, content-strategy

Cambios pedidos por el creador / pendientes

Sin pendientes en el checklist ✅

sentiment del cliente → Ninjo · caché (2026-07-24 15:13:31) · modelo claude-sonnet-4-6 · ventana 7d
+1
Sentiment score
conforme_activo
Bajo
Churn risk
del lado del cliente
0
Cambios pedidos
pendientes en checklist
0
Volumen cliente
agente: 2 convs

Matriz de volumen

re chequear
Cliente callado y agente con poco tráfico: ambiguo (puede ser período tranquilo). Re-chequear próxima ventana.

Sentiment del cliente

📋 Sin citas textuales del cliente disponibles en los digests (todo el contenido es narración del equipo Ninjo), por lo que no se puede confirmar sentimiento explícito; sin embargo, el patrón conductual indica un cliente activo, colaborativo y sin señales de churn. Alerta puntual: caída del -90% en volumen de conversaciones semana del 6-12 jul requiere diagnóstico urgente antes de que afecte la percepción del cliente sobre el ROI de la plataforma. Acción CS recomendada: (1) presentar análisis de causa raíz de la caída de volumen esta semana con plan de remediación concreto; (2) solicitar a Joe feedback explícito sobre satisfacción para generar cita real y mejorar la señal de sentimiento.
SeñalValorEvidencia (cita del cliente)
Enojado con la plataformano
Es quejano
Queja constructiva (accionable)no

Evidencia citada

SeñalCita textualFecha
Participación activa y sostenida en iteraciones del agente — Joe co-construye el prompt semana a semana“None”2026-05-19
Joe provee su definición exacta de adaptógenos para incorporar al agente — nivel de ownership alto“None”2026-06-08
Joe delega campaña publicitaria y coordina call de verificación con Lolo — colaboración operativa fluida“None”2026-05-26
Joe acepta escalar issues vía tag @Ninjo — adopta el workflow propuesto por el equipo“None”2026-06-16
Joe comprometido a hacer outreach manual a 135 leads — ejecuta recomendaciones del equipo“None”2026-06-10

Cambios pedidos por el creador (Métrica 1)

Sin cambios pendientes en el checklist ✅