opensource/n8n
3
0
Fork 0
mirror of https://github.com/n8n-io/n8n.git synced 2026-10-24 02:09:26 +08:00
Code Issues Packages Projects Releases Wiki Activity

chore(ai-builder): Point the Agent model steps at the resolve_llm next step (no-changelog)

Browse Source 
agent_builder_resolve_llm now returns what to do with each result. The
agent-builder skill no longer repeats that guidance.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
Mutasem Aldmour
2026-10-23 18:25:17 +02:00
co-authored by Claude Opus 5.5
parent cc33091577
commit 5baaf964c7
1 changed files with 5 additions and 16 deletions
Download Patch File Download Diff File Expand all files Collapse all files
packages/@n8n/instance-ai/skills/agent-builder/SKILL.md
+5 -16
View File
@@ -424,13 +424,8 @@ Use this to resolve the target Agent's main `model` and `credential`.
with provider and model when the user named them, otherwise with no
arguments.
2. If `agent_builder_resolve_llm` succeeds, persist `model = "{provider}/{model}"`
and `credential = credentialId`. If the result has
`claimedFreeOpenAiCredits: true`, tell the user that you set them up with
free OpenAI credits. If the result's `credentialName` is "Gateway credits",
persist it like any credential and tell the user that the model runs on
Gateway credits. Never show the internal "n8n Connect" or "AI Gateway"
names. If it has `autoPicked: true`, tell the user which provider and model
you picked and that they can ask to change it. Do not ask for confirmation.
and `credential = credentialId`, then follow the result's `nextStep`. Never
show the internal "n8n Connect" or "AI Gateway" names.
3. If the user asks to pick, change, confirm, or configure a model or main
credential, ask with `agent_builder_ask_questions`.
- Exception: when the user explicitly asks to use Gateway credits for the
@@ -455,15 +450,9 @@ Use this to resolve the target Agent's main `model` and `credential`.
user can set the model in the panel. For a model change on an existing
Agent, ask immediately instead, and never write `model: ""` over an existing
model.
5. If `agent_builder_resolve_llm` reports `unknown_model`, retry with a value
from the returned `availableModels` or ask with
`agent_builder_ask_questions`. This can also occur when you call it with no
model. The returned `availableModels` is the live list of the credential, so
pick from it. Do not guess, and do not reuse the model id that you just
tried. When `availableModelsTruncated: true`, the
list is only a sample, so ask instead of assuming that the model is absent.
If it reports `model_lookup_failed`, retry `agent_builder_resolve_llm`. Do
not guess a model.
5. If `agent_builder_resolve_llm` reports `unknown_model`, follow its
`nextStep`. If it reports `model_lookup_failed`, retry
`agent_builder_resolve_llm`. Do not guess a model.
6. If `agent_builder_call_agent` fails with `code: "invalid_model"`, the
provider rejected the model id, not the credential. Do not tell the user to
check their API key. Call `agent_builder_resolve_llm` for that provider