evaluar-contexto-necesario

name	evaluar-contexto-necesario
description	Use when preparing a complex task for an AI agent. Evaluates what files, examples, or templates should be attached as context to get the best result. Invoke with: /evaluar-contexto-necesario, what context do I need, prepare context, context for agent, attach files.
argument-hint	Describe the task you are about to run (e.g. 'train a classifier on the Iris dataset')
user-invocable	true

Skill: evaluar-contexto-necesario

Evaluate and recommend the context that should be attached to a complex AI agent task.

Steps

1. Parse the Task

• Restate the task in one sentence.
• Identify: what code will be created or modified, what data will be used, what output is expected.

2. Apply the Context Checklist

Work through references/context-checklist.md for the task.

3. Recommend Files to Attach

• List specific files from the project that the agent will need.
• Prioritise: spec > existing implementation > tests > examples.
• Explain why each file is needed (not just list them).

4. Warn About Context Size

• If the total estimated tokens of recommended files exceeds ~8 000, flag it.
• Suggest pruning: "You need only the train() function from train_model.py, not the whole file."

5. Flag Missing Context

• If the task requires information that does not exist yet (a spec, a schema, an example), say so explicitly: "You should first create doc/spec.md using the spec-a-tareas skill."

Output

A prioritised list of files to attach with justifications, plus a warning if context would be too large.