dotnet-toolkit
AGENTS.md2 starsChanged 53 days agoMITView on GitHub
CLAUDE.md is this repository's canonical operating contract. Read and follow it; this file intentionally does not duplicate those instructions. Before the first C# read or edit, also read .claude/rules/dotnet-index.md. Claude Code loads that router automatically; Codex does not scan .claude/rules/. The Markdown files under agents/ are the canonical role briefs, and the adapter below is the supported way to run them outside Claude Code — there are deliberately no generated .codex/agents/*.toml equivalents to drift from them. When a skill asks…
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 31,843 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 repository10 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- dotnet-architectureSkill
- dotnet-consistencySkill
- dotnet-exploreSkill
- dotnet-initSkill
- dotnet-performanceSkill
- dotnet-readSkill
- dotnet-reviewSkill
- dotnet-selfevalSkill
- dotnet-writeSkill
The file
We split this AGENTS.md into the rules it gives an agent
1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 6
Before the first C# read or edit, also read .claude/rules/dotnet-index.md.
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 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.
Files
Codex guidance
CLAUDE.md is this repository's canonical operating contract. Read and follow it; this
file intentionally does not duplicate those instructions.
Before the first C# read or edit, also read .claude/rules/dotnet-index.md.
Claude Code loads that router automatically; Codex does not scan .claude/rules/.
The Markdown files under agents/ are the canonical role briefs, and the adapter below is
the supported way to run them outside Claude Code — there are deliberately no generated
.codex/agents/*.toml equivalents to drift from them. When a skill asks for a named agent and that
custom agent is unavailable in the current host, launch the closest built-in subagent and tell it to
read the matching agents/<name>.md before starting. Treat the YAML frontmatter fields that name
Claude-specific tools, models, memory, or colors as host configuration; the Markdown body remains the
shared role contract.
That frontmatter's tools: list is host configuration in one further respect: it spells every MCP
tool with Claude Code's plugin-qualified prefix, mcp__plugin_dotnet-toolkit_dotnet__<tool>. Codex
registers the same server under the bare name dotnet and namespaces by that instead. Resolve the
prefix once by probing for one known tool rather than assuming either form, and apply what comes back
to the rest — a qualified name that misses means the label differs, never that the tools are absent.
Checking the installation
codex mcp list --json reports configuration, not health: it launches nothing, so it shows the
server enabled whether or not the command it names can run. To verify for real —
executor, effective CODEX_HOME, whether each stored path is openable from this process, the SDK the
workspace loads with, and a live handshake — run:
dotnet dist/DotnetToolkit.McpServer.dll doctor --repo .It exits nonzero on any failure. doctor --no-probe additionally fails when dist/ has fallen
behind src/, which is the pre-commit gate for the rule that dist/ is committed alongside the
change that required it.
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
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.

