From 645c38d43c682ba7e383d354c06c301190e03c68 Mon Sep 17 00:00:00 2001 From: JB Date: Wed, 26 Aug 2026 19:14:06 -0400 Subject: [PATCH] chore(agent): refine global behavior guidance --- docs/global.AGENTS.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/global.AGENTS.md b/docs/global.AGENTS.md index 8ad3b20..4d941e1 100644 --- a/docs/global.AGENTS.md +++ b/docs/global.AGENTS.md @@ -78,7 +78,7 @@ Below you will find my preferences so that we can be more aligned as we work tog - ADRs should be functionally immutable, but not read-only. If a new decision invalidates a previous decisions, update that ADR with a link to the new ADR. - Follow the guidelines in the [`writing-for-agents`](../skills/document/writing-for-agents/SKILL.md) skill when creating documentation. -## Visual & Design Work +## Visual & Design Work - Do not edit real components first. If the work is non-trivial UI changes, layout, or copy, build a few distinct static mocks, and ready them for manual review and wait for my feedback before implementing. - I prefer dark mode first. Minimal copy. Minimal chrome. No em dashes. Don't add meaningless text and measurements, everything should have a purpose. @@ -94,9 +94,9 @@ Below you will find my preferences so that we can be more aligned as we work tog - Prototype or create PoCs for large features first. If we can identify if something will or will not work early, that will save time in the long run. - Do not touch production or live databases unless explicitly instructed to do so. - When presenting any single local file for my review, or a structured long-form report, plan, ADR, review, or summary, load the [`artifact-presenter`](../skills/document/artifact-presenter/SKILL.md) skill and open the complete file with mcat in its reusable Herdr pane instead of returning the full artifact inline. +- After reporting status, if there is no input required from me, continue work on the task. ## Subagents and Parallel work -- Not every ask needs a subagent. Do not spawn up subagents or multi-agent workflows for something a single agent can finish in a single pass. Delegate appropriately for adversarial review, or specialized parallel work. - When we have several agents working together on a feature, state file ownership up front so they do not collide.