agentleFS
MicrosoftKnown publishermicrosoft/mxc/AGENTS.md

mxc

AGENTS.md2.9k starsChanged 12 days agoMITView on GitHub

Read and follow .github/copilot-instructions.md before writing or reviewing code. It contains MXC's architecture, compatibility, security, build, and SDK rules. The expectations below apply in addition. - Understand the existing flow before editing: callers, shared helpers, backend ownership, errors, cleanup, and affected SDK entry points. A working happy path is not enough; check how the change composes with existing behavior. - Put behavior where it belongs. Parsers decode contracts and preserve field presence; backend-specific defaults, policy checks, and enforcement belong in…

We split this AGENTS.md into the rules it gives an agent

11 rules. About 0 are common, 11 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈11

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Read and follow .github/copilot-instructions.md before writing or reviewing code.

    line 3
  2. Distinctive:

    Put behavior where it belongs.

    line 13
  3. Distinctive:

    If the design cannot express the feature cleanly, refactor the responsible component while preserving existing callers and behavior.

    line 19
  4. Distinctive:

    Consider whether a maintainer can understand and safely extend the result.

    line 26
  5. Distinctive:

    If adding a feature requires replacing or deleting a large part of a subsystem, split the work into reviewable vertical slices: a small …

    line 30
  6. Distinctive:

    Use plain, precise language in code reviews, comments, documentation, and handoffs.

    line 38
  7. Distinctive:

    Keep comments sparse and short: explain a non-obvious reason or constraint, not what the next line does.

    line 43
  8. Distinctive:

    Update the relevant canonical document rather than adding overlapping Markdown.

    line 48
  9. Distinctive:

    Document only what the reader needs: supported behavior, usage, prerequisites, limitations, and useful error remedies.

    line 49
  10. Distinctive:

    Prefer tests of useful behavior, feature interactions, and failure handling over many tiny checks of implementation details.

    line 66
  11. Distinctive:

    Follow the validation runbook and backend guide.

    line 74

≈ “Common” is an estimate: at least 1 in 5 of the 36,368 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
AGENTS.md

MXC agent instructions

Read and follow .github/copilot-instructions.md before writing or reviewing code. It contains MXC's architecture, compatibility, security, build, and SDK rules. The expectations below apply in addition.

Design for the codebase, not just the feature

  • Understand the existing flow before editing: callers, shared helpers, backend ownership, errors, cleanup, and affected SDK entry points. A working happy path is not enough; check how the change composes with existing behavior.
  • Put behavior where it belongs. Parsers decode contracts and preserve field presence; backend-specific defaults, policy checks, and enforcement belong in the backend, not shared parsing code.
  • Reuse the normal implementation path. Do not add a parallel helper, a feature-specific variant, or a narrow exception just to make a feature work. If the design cannot express the feature cleanly, refactor the responsible component while preserving existing callers and behavior.
  • Choose the simplest design that fits the current requirements. Add an abstraction only when it solves a real problem; do not build for imagined reuse. Name functions for their responsibility, not a growing list of cases.
  • Consider whether a maintainer can understand and safely extend the result. Completing the requested feature does not excuse duplication, misplaced responsibilities, weakened checks, or hidden failure paths.
  • If adding a feature requires replacing or deleting a large part of a subsystem, split the work into reviewable vertical slices: a small part of the feature working through the relevant layers, with its tests. Separate prerequisite refactoring from behavior changes; keep each step buildable and preserve existing behavior. Do not attempt the whole rewrite at once.

Write for everyday developers

Use plain, precise language in code reviews, comments, documentation, and handoffs. Explain the concrete behavior and consequence; avoid invented jargon and research-paper language. Be concise without leaving out the detail needed to understand or act. Neither verbosity nor oversimplification is helpful.

Keep comments sparse and short: explain a non-obvious reason or constraint, not what the next line does. Prefer clear names and straightforward code.

Documentation should describe the finished design, behavior, and usage, not an agent's thought process, implementation diary, or discarded approaches. Update the relevant canonical document rather than adding overlapping Markdown. Document only what the reader needs: supported behavior, usage, prerequisites, limitations, and useful error remedies. Include implementation detail only when it explains a necessary maintenance or design decision, in developer docs.

Separate paragraph-length Markdown list items with a blank line; keep short lists compact.

Require meaningful behavior coverage

  • For a feature, add or extend a functional validation test in the applicable CI suite. Backend behavior must run in scheduled/nightly validation on the relevant hosts; adding a test file without wiring it into CI is not enough.
  • For a bug fix, add a regression scenario that exercises the reported failure through the relevant public API or backend flow and checks the correct result. Where practical, show that it fails before the fix and passes afterward.
  • Prefer tests of useful behavior, feature interactions, and failure handling over many tiny checks of implementation details. Extend existing scenarios rather than duplicating them.
  • Unit tests, including tests at the bottom of .rs files, are useful for isolated logic but do not replace functional validation. Do not judge coverage by test count or tests that merely repeat the new implementation.
  • Follow the validation runbook and backend guide. Check that the intended scenario actually runs and asserts behavior on a suitable host. State what was not run or cannot be covered; do not invent evidence. Documentation-only changes do not need runtime tests.

For reviews, use the MXC review skill. Focus on high- and medium-severity defects and consolidate findings into one review, rather than creating extra review rounds for nits or optional cleanup.

Download now (ZIP)

Trending

Steady: 0 new projects took it in the last 4 weeks, none the 4 weeks before.thin data

Week of Aug 30: 0 new projectsWeek of Sep 6: 0 new projectsWeek of Sep 13: 0 new projectsWeek of Sep 20: 0 new projectsWeek of Sep 27: 0 new projectsWeek of Oct 4: 0 new projectsWeek of Oct 11: 0 new projectsWeek of Oct 18: 0 new projects
Aug 2026Sep 2026Oct 2026
New projects taking a version of this file each week, over the 8 weeks our latest run read: only 0 in all, so this is a hint, not a trend.

Who's talking about microsoft/mxc

The biggest 5 of 66 mentions linking to the repository that our crawl found. Each line is their own headline.

See all 66

Talked about on GitHub, the web, Bluesky and Hacker News (66 times).

On GitHub · 39

34 more on GitHub

On the web · 14

9 more on the web

On Bluesky · 10

5 more on Bluesky

On Hacker News · 3

  • Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.

    Of the AGENTS.md files we read, the 28,693 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.

  • In this repository3 other files it gives its agents.

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.