docs: restructure AI guidance for progressive disclosure - #966
Open
hypnotoad08 wants to merge 1 commit into
Open
hypnotoad08 wants to merge 1 commit into
hypnotoad08 wants to merge 1 commit into
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Restructures the repository's AI/agent guidance from ~20 overlapping, tool-specific files into a
small set of docs organized around progressive disclosure.
The idea: give an agent a short entry point plus a map, and let it pull in deeper detail only for
the area it's actually touching, instead of front-loading everything for every task.
The old set loaded the same guidance five-plus times under different vendor filenames
(
ANTHROPIC.md,BEDROCK.md,BARD.md,COHERE.md,AZURE_OPENAI.md,HUGGINGFACE.md,OpenAI.md), split further across per-editor rule files (.cursorrules,clinerules,windsurfrules). An agent reading one of these had no way to know which of the ~20 filesactually applied to it, and a task touching one subsystem (e.g. linting) still had to wade
through unrelated content to find the relevant paragraph. Routing by concern instead of by tool
means each doc only surfaces when it's relevant to what's being touched.
Changes
Added
AGENTS.md: concise root entrypoint with project overview, layering rules, and atable routing to deeper docs by task ("touching persistence? read X"; "touching frontend?
read Y").
fe/AGENTS.md: scoped frontend conventions (Composition API, Pinia,SignalR/HTTP state handling, a11y, test commands).
SECURITY_RULES.md: consolidated secure-coding guidance (OWASP/CWE),separate from
SECURITY.md's vulnerability-reporting policy.Changed
.github/copilot-instructions.md: trimmed to a compatibilityentrypoint that points to
AGENTS.mdand the other root docs, so Copilot-specific guidancedoesn't drift from the same rules everyone else reads.
BACKEND_ARCHITECTURE.mdandLINTING.md: updated asthe linked-to detail for their areas from
AGENTS.md.Removed
~2,170 lines of duplicated/per-vendor instruction files:
ANTHROPIC.md,BEDROCK.md,BARD.md,COHERE.md,AZURE_OPENAI.md,HUGGINGFACE.md,OpenAI.md,.cursorrules,clinerules,windsurfrules,WARP.md,.github/CLAUDE.md,.github/AGENTS.md,.github/RULES.md,.github/CONVENTIONS.md, and the strayvuejs.md/.mdc/NET_ASP.NET Core_rules.mdcfiles. Each one said roughly the same thing to a different tool. The uniquereview, testing, linting, and compatibility rules they carried are preserved, folded into the
docs above rather than dropped.
The provider files (
ANTHROPIC.md,BEDROCK.md,BARD.md,COHERE.md,AZURE_OPENAI.md,HUGGINGFACE.md,OpenAI.md) were never auto-discovered by anything; they were hand-linkedreference docs with duplicate boilerplate, so there's no tool-side loss from removing them.
The per-editor files are a different case, and I checked each one against current docs before
removing it:
AGENTS.mdnatively as a documented alternative to.cursor/rules.AGENTS.mdalongside.clinerules/, shown together in its Rules panel.AGENTS.mdper its current (Cognition/Devin-owned) docs;.windsurf/rulesstill works as the legacy path.AGENTS.mdis now the recommended default, withWARP.mdstillsupported and only taking priority if both files exist in the same directory. We deleted
WARP.mdoutright rather than keeping a thin pointer, which is intentional here: it makesAGENTS.mdunambiguously authoritative instead of leaving two files Warp could pick between.Testing
Documentation-only change; no application tests run.
rules (security, testing, linting, compatibility) were carried forward, not dropped.
AGENTS.md,fe/AGENTS.md,copilot-instructions.md,SECURITY_RULES.md).discover
AGENTS.mdwithout needing their previous vendor-specific filename.Notes
If any contributor still uses a tool that hasn't picked up
AGENTS.mdsupport yet, its old rulesfile can be restored as a thin pointer to
AGENTS.md(same pattern as.github/copilot-instructions.md)