n8n-architect skill description suppresses auto-invocation ("explicitly"), and never mentions .workflow.ts

[tmdrtmdr]

File: packages/skills/src/agent-skills/n8n-architect/SKILL.md (observed on v2.3.1)

Current description:

Use when the user explicitly wants to create, edit, validate, sync, or troubleshoot n8n workflows, asks about n8n nodes or automation, or wants to use n8n-as-code in the current context root.

Two problems with it:

1. "explicitly" reads to an agent as do not load this unless asked for by name. It suppresses exactly the automatic invocation the skill is designed for.
2. The description never mentions .workflow.ts, so an agent opening the file format this project invented has no trigger to load the skill that documents it.

Measured

Parsing 30 days of local agent transcripts for actual skill invocations (not mentions), on one machine:

• n8n-architect was invoked 3 times in 30 days.
• Over the same window, 18 separate agent sessions ran n8nac push/pull operations.
• The six most active workflow-editing sessions invoked it 0, 1, 1, 0, 0, 0 times.

Invocation is inversely correlated with how much workflow editing a session actually does.

That is despite our own project context file carrying an explicit line reading "Invoke the n8n-architect skill before editing." — in a file that autoloads in every session in that folder. An instruction to load a skill loses to the agent's sense that it already knows enough, so the description has to do the work.

Suggested

Use for any task that reads, writes, or debugs an n8n workflow — including editing a .workflow.ts file, choosing node parameters, or syncing with an n8n instance.

Happy to PR it.