find-standards ยท diff

git:20260817.b145775 to git:20260817.034c32a

8 added, 4 removed. Audit A to A.

What this update changes in the checks

No check changed: every fixed check gives the same answer before and after this update.

A tripped check is a label, not a verdict: read the matched text and the lines below. This week's updates that newly trip a check.

---
name: find-standards
description: Find how this problem is already solved: standards we could adopt, and the industry's best practices.
disable-model-invocation: true
---
- Our #1 goal when desiging software is to fit into the existing ecosystem.
+ Our #1 goal when desiging software is to make intelligent decisions when building a feature around:
+ - Copying/adapting existing patterns for code, design, domain, dependencies, and/or infrastructure
+ - Adding new patterns for code, design, domain, dependencies, and/or infrastructure.
+ - Recognizing existing pattern is outdated or bad design or just no longer works for us and so we decide to refactor while copying/adapting/adding.
+
Find how the problem in front of the user is already solved, and bring back two things:
- - **Standards we could adopt (Internal)**: a library, a (existing or potential) shared service, an existing dependency, an existing module. Re-using what we've already done and keeping our stack small cuts duplication and complexity. What do we already do that is applicable to this problem?
+ - **Standards we could adopt (Internal)**: In order to capitlize on what we've already built by keeping our stack, domain, design, and codebase small we can avoid duplication and complexity. What do we already do that is applicable to this problem?
- **Best practices (External)**: how the industry solves this, and the problem each practice solves.
- Fan out subagents in one batch, one per angle, and cover both sources below. Each subagent starts blank, so give it a plain description of the problem, our stack, and the constraints. The subagents read; you judge what comes back, and report.
+ Fan out subagents in one batch, one per angle, and cover both sources below and receive an extremely detailed response (written by /writing-for-agents). Each subagent starts blank, so give it a description of the problem and what its trying to find. The subagents read; you judge what comes back, and report.
- **Internal**: code we already own. Search this codebase first. Search the rest of the GitHub org too if this project exists within an org or company.
- - **External**: libraries, standards, the pattern with a name, the RFC, the write-up by the team who hit this at scale. Split the angles by question, not by search engine: what people build with, what they name it, what they regret.
+ - **External**: libraries, standards, best practices, the pattern with a name, the RFC, the write-up by the team who hit this at scale. Split the angles by question, not by search engine: what people build with, what they name it, what they regret. Try to find evidence/proof and not just opinion.
Every standard names a real artifact: a library, a file path with lines, a spec, a doc URL. "Use a state machine" is not a finding; `xstate` and `src/order/machine.ts:40` are. Every practice names the problem it solves, so we can tell whether we have that problem.
Report one table: each candidate, what it is, where it is already in use, and what it costs us to take. Show us both the internal and external for each candidate (leave blank if nothing was reported). Then give your recommended default, and the one fact that would change it.
Done when both sources have reported, and each candidate carries a verdict: **adopt**, **adapt**, or **ruled out** with the reason.