Scrum Daily Update Generator

Transforms a technical conversation or work session into a Daily Scrum update formatted for fast, clear team communication. Output is designed to be spoken aloud in under 30 seconds.

Step-by-step process

1. Analyze the conversation

Scan the full conversation and extract:

Tasks worked on — features, bugs, refactors, research, integrations, reviews
Decisions made — architectural choices, library selections, approach changes
Solutions implemented — actual code written, problems solved, PRs opened
Blockers or open questions — unresolved dependencies, unclear requirements, waiting on someone
Planned next steps — things explicitly mentioned as "next" or implied by incomplete work

If the conversation is very long or covers multiple topics, group by feature/task area before summarizing.

2. Translate technical details into team-readable bullets

Apply these translation rules:

Instead of...	Write...
"Fixed the NullPointerException in UserService.java line 247"	"Fixed a crash in the user service affecting login flow"
"Refactored the Redux store to use RTK slices"	"Cleaned up state management in the frontend"
"Investigated S3 presigned URL expiry issue with boto3"	"Investigated file upload expiry bug in cloud storage"
"Spent 2h reading Next.js docs on ISR"	"Researched rendering strategy options for the dashboard"

Rule: Strip implementation noise. Keep business/feature context.

3. Generate the daily update

Use the standard format below. Choose between Spanish or English based on the user's language.

Output format

Standard (recommended for most teams)

### Daily Update

**Ayer trabajé en**
- [bullet 1 — main task or feature]
- [bullet 2 — secondary task, bug, or investigation]
- [bullet 3 — optional: decision, review, or research]

**Hoy planeo**
- [bullet 1 — clear next action]
- [bullet 2 — optional: follow-up or dependent task]

**Bloqueos**
- [Describe blocker clearly, or: ninguno]

English version (for international/mixed teams)

### Daily Update

**Progress**
- [bullet 1]
- [bullet 2]

**Next**
- [bullet 1]
- [bullet 2]

**Blockers**
- [blocker or: none]

Extended version (when impact context adds value)

Add an Impacto / Impact section after blockers when the work has clear team or product value:

**Impacto**
- [One line: what this enables or unblocks for the team]

Formatting rules

Maximum 3–4 bullets per section — ruthlessly prioritize
Each bullet = 1 clear sentence — no sub-bullets, no parentheticals
Start bullets with a verb — Implemented, Fixed, Investigated, Reviewed, Defined, Deployed
No jargon without context — translate technical terms into feature/outcome language
Blockers must be actionable — name what you need, who you're waiting on, or what's unclear
"Ninguno / None" is always valid for blockers — don't invent friction
Hoy/Next section — only include things you can realistically start today

Multi-task grouping (for long or complex sessions)

If the session covered 3+ distinct areas, group before summarizing:

Area 1: Auth service refactor → 2 bullets
Area 2: Bug investigation → 1 bullet
Area 3: Code review → 1 bullet

Then merge into the most impactful 3–4 bullets for the final output. When in doubt, lead with the item that most directly affects the sprint goal.

Example outputs

Example 1 — Feature implementation session

Context: Session where a developer built a skill for exporting AI conversation summaries.

### Daily Update

**Ayer trabajé en**
- Implementación del skill para generar resúmenes de trabajo desde conversaciones con IA.
- Diseño de estructura reutilizable para skills de agentes en el repositorio del equipo.
- Definición del formato estándar para exportar avances desde sesiones de trabajo.

**Hoy planeo**
- Integrar el skill dentro del repositorio de AI skills del equipo.
- Documentar el formato de output para que otros devs puedan adaptarlo.

**Bloqueos**
- Ninguno.

**Impacto**
- Reduce el tiempo de preparación del daily para quienes trabajan con agentes IA.

Example 2 — Debugging session

Context: Session investigating a slow API endpoint.

### Daily Update

**Progress**
- Identified the root cause of slow response times on the /orders endpoint (N+1 query).
- Implemented eager loading fix and verified a 4x improvement in local tests.

**Next**
- Open PR for the query fix and request review.
- Monitor query performance in staging after merge.

**Blockers**
- Need DB access in staging to run EXPLAIN ANALYZE — waiting on DevOps.

Example 3 — Research / investigation session

Context: Session evaluating auth libraries for a new project.

### Daily Update

**Ayer trabajé en**
- Investigación de opciones de autenticación para el nuevo servicio (Auth0, Cognito, Keycloak).
- Comparación de costos, integración con nuestro stack y curva de aprendizaje del equipo.
- Decisión preliminar: Auth0 como opción principal, Keycloak como backup self-hosted.

**Hoy planeo**
- Presentar resumen de opciones al equipo en la sesión de arquitectura.
- Iniciar spike técnico con Auth0 si hay consenso.

**Bloqueos**
- Ninguno por ahora, pero la decisión final depende del presupuesto aprobado por producto.

Tone calibration

Team type	Tone	Example
Junior/mixed	Descriptive, clear	"Arreglé el bug de login que bloqueaba a usuarios nuevos"
Senior/technical	Concise, outcome-first	"Fixed auth regression in new user flow"
Cross-functional	Business-outcome focus	"Resolved issue preventing new user onboarding"

Default to descriptive and clear unless the conversation context suggests a senior technical team.

When context is ambiguous

If the conversation doesn't clearly indicate what was "yesterday" vs "today", use this heuristic:

Completed work / resolved problems → Ayer / Progress
Partially done / still in progress → split across both sections
Explicitly mentioned as next step → Hoy / Next
Mentioned but stuck → Bloqueos / Blockers

If the session is too short or lacks enough detail, ask the user: "Do you want to add anything that wasn't covered in the conversation?"

Reference Guide

Load the corresponding reference file based on the generation need:

Reference	Purpose	When to load
`references/output-examples.md`	Complete output examples by complexity level (trivial, standard, complex, multi-feature) and by team type (junior, senior, cross-functional)	Use as calibration targets when generating any update
`references/translation-rules.md`	Rules for translating implementation details into team-readable language, organized by domain (frontend, backend, infra, data, auth, testing)	Use when converting technical conversation content into standup bullets

scrum-daily-update-generator