mirror of
https://github.com/humanlayer/12-factor-agents.git
synced 2025-08-20 18:59:53 +03:00
504ad6f544
- Update all factor file names to use two-digit numbering (e.g., factor-01, factor-02) - Update all internal links to reflect new file naming convention - Update image references to match new naming pattern - Maintain consistent navigation links across all factor documents This change improves file organization and makes the documentation structure more maintainable.
3.8 KiB
3.8 KiB
Factor 13 - pre-fetch all the context you might need
If there's a high chance that your model will call tool X, don't waste token round trips telling the model to fetch it, that is, instead of a pseudo-prompt like:
When looking at deployments, you will likely want to fetch the list of published git tags,
so you can use it to deploy to prod.
Here's what happened so far:
{{ thread.events }}
What's the next step?
Answer in JSON format with one of the following intents:
{
intent: 'deploy_backend_to_prod',
tag: string
} OR {
intent: 'list_git_tags'
} OR {
intent: 'done_for_now',
message: string
}
and your code looks like
thread = {"events": [initial_message]}
next_step = await determine_next_step(thread)
while True:
switch next_step.intent:
case 'list_git_tags':
tags = await fetch_git_tags()
thread["events"].append({
type: 'list_git_tags',
data: tags,
})
case 'deploy_backend_to_prod':
deploy_result = await deploy_backend_to_prod(next_step.data.tag)
thread["events"].append({
"type": 'deploy_backend_to_prod',
"data": deploy_result,
})
case 'done_for_now':
await notify_human(next_step.message)
break
# ...
You might as well just fetch the tags and include them in the context window, like:
- When looking at deployments, you will likely want to fetch the list of published git tags,
- so you can use it to deploy to prod.
+ The current git tags are:
+ {{ git_tags }}
Here's what happened so far:
{{ thread.events }}
What's the next step?
Answer in JSON format with one of the following intents:
{
intent: 'deploy_backend_to_prod',
tag: string
- } OR {
- intent: 'list_git_tags'
} OR {
intent: 'done_for_now',
message: string
}
and your code looks like
thread = {"events": [initial_message]}
+ git_tags = await fetch_git_tags()
- next_step = await determine_next_step(thread)
+ next_step = await determine_next_step(thread, git_tags)
while True:
switch next_step.intent:
- case 'list_git_tags':
- tags = await fetch_git_tags()
- thread["events"].append({
- type: 'list_git_tags',
- data: tags,
- })
case 'deploy_backend_to_prod':
deploy_result = await deploy_backend_to_prod(next_step.data.tag)
thread["events"].append({
"type": 'deploy_backend_to_prod',
"data": deploy_result,
})
case 'done_for_now':
await notify_human(next_step.message)
break
# ...
or even just include the tags in the thread and remove the specific parameter from your prompt template:
thread = {"events": [initial_message]}
+ # add the request
+ thread["events"].append({
+ "type": 'list_git_tags',
+ })
git_tags = await fetch_git_tags()
+ # add the result
+ thread["events"].append({
+ "type": 'list_git_tags_result',
+ "data": git_tags,
+ })
- next_step = await determine_next_step(thread, git_tags)
+ next_step = await determine_next_step(thread)
while True:
switch next_step.intent:
case 'deploy_backend_to_prod':
deploy_result = await deploy_backend_to_prod(next_step.data.tag)
thread["events"].append(deploy_result)
case 'done_for_now':
await notify_human(next_step.message)
break
# ...
Overall:
If you already know what tools you'll want the model to call, just call them DETERMINISTICALLY and let the model do the hard part of figuring out how to use their outputs
Again, AI engineering is all about Context Engineering.