Agents
AGENTS.md7 starsChanged 2 months agoMITView on GitHub
Always active; routes to the topic files required by the current task.
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 31,774 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 repository29 other files it gives its agents.
- address-pr-commentsSkill
- build-and-scan-container-imageSkill
- execute-database-migrationSkill
- generate-adrSkill
- generate-service-documentationSkill
- get-pr-commentsSkill
- incident-triageSkill
- install-dev-toolsSkill
- prepare-releaseSkill
- remediate-dependency-vulnerabilitiesSkill
- review-configurationSkill
- twg-agentic-searchSkill
- twg-artifactsSkill
- twg-code-reviewSkill
- twg-confluenceSkill
- twg-context-discoverySkill
- twg-engineering-workSkill
- twg-jira-resolve-merged-workSkill
- twg-jiraSkill
- twg-operational-healthSkill
- twg-responsibility-routingSkill
- twgSkill
- twg-space-creationSkill
- twg-status-rollupsSkill
- validate-and-deploy-iacSkill
- validate-and-deploy-kubernetesSkill
- validate-documentationSkill
- validate-github-actionsSkill
- verify-service-healthSkill
The file
We split this AGENTS.md into the rules it gives an agent
9 rules. About 0 are common, 7 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈7
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 72
When two files conflict, the more specific file wins.
- Distinctive: line 96
Don't add explanatory preamble before code ("Here is a function that...").
- Distinctive: line 97
Don't apologize or over-explain errors.
- Distinctive: line 98
When multiple approaches exist, recommend one and explain the trade-off briefly — don't list every option.
- Very specific: line 99
Never add TODO comments unless explicitly asked.
- Very specific: line 100
Never generate placeholder logic (e.g., // implement later) — implement it or say you can't.
- Distinctive: line 101
Prefer editing existing code over rewriting files wholesale.
- Distinctive: line 102
When unsure about intent, ask one targeted question rather than guessing.
- Distinctive: line 108
When calling command line tools, add a comment at the end of the command line explaining the purpose of the command.
≈ “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
Agent Guidance
This file is always active. Apply its rules to every task, then load only the topic files required by the task. Do not load every topic file by default.
Topic Routing
Load the matching file before making decisions or edits in that topic:
| Topic | File to load |
| Git repositories, branches, commits, or pull requests | git.md |
| Source code, scripts, or general coding practice | coding.md |
| Markdown files or documentation authoring | markdown.md |
| Security engineering or supply-chain security | security.md |
| Test strategy or test implementation | testing.md |
| Container images or container runtime | containers.md |
| Logs, metrics, traces, or service health | observability.md |
| Releases, deployments, or change promotion | release.md |
| Production incidents or operational response | incident-response.md |
| Software architecture or design decisions | architecture.md |
| Dependency lifecycle or software supply chain | dependencies.md |
| API design or service contracts | api.md |
| Databases, schemas, or data migrations | database.md |
| Performance or capacity engineering | performance.md |
| Kubernetes workloads | kubernetes.md |
| Cloud architecture or cloud operations | cloud.md |
| Disaster recovery or business continuity | disaster-recovery.md |
| Cloud cost management or FinOps | cost-management.md |
| Team engineering workflow | engineering-workflow.md |
| Compliance, risk, or audit controls | risk-and-compliance.md |
| Service ownership or operational documentation | service-ownership.md |
| Accessibility for user-facing software | accessibility.md |
| Application configuration, settings, or secrets | configuration.md |
| CMake build-system selection | cmake_make.md |
| GNU Makefiles and task runners | makefile.md |
| GitHub Actions workflows | githubactions.md |
| Go | go.md |
| Bash, Zsh, or shell scripts | shell.md |
| PowerShell | powershell.md |
| C# or .NET | dotnet.md |
| JavaScript or TypeScript | typescript.md |
| HTML | html.md |
| Rust | rust.md |
| Rust style, API, and lint conventions | rust_style_guide.md |
| SSL/TLS connections or corporate proxy certificates | ssl-tls.md |
| Terraform | terraform.md |
| Tauri applications | tauri.md |
Delegating work to local LM Studio models, or the lmstudio-rs-mcp server | lmstudio-subagents.md |
| Terminal user interfaces | tui.md |
| Common User Access TUI specification | CUA_TUI_Specification.md |
| Twelve-Factor application design | twelvefactors.md |
| GUI automation for macOS, Windows, or Linux | gui_automation.md |
| Terminal.Gui theme subsystem (C#) | terminal_gui_themes.md |
| Code snippets | snippets/SNIPPETS.md |
| Skills (on-demand workflows) | skills/<name>/SKILL.md; see the Skills section of README.md |
Load more than one file when a task spans topics. For example, a Tauri
application with a Rust backend and Terraform deployment requires
tauri.md, rust.md, and terraform.md.
Guidance Precedence
Apply this file to every task. Load every matching topic file before making decisions or edits.
When two files conflict, the more specific file wins. From highest to lowest:
- Language files (
rust.md,go.md, and so on) and framework files coding.md- Other topic files
- The general guidance in this file
Exceptions:
- General Behavior and Tool Calling below govern the agent's own conduct and are never overridden by a topic file.
CUA_TUI_Specification.mdoverridestui.mdwhen it applies.- A project's own conventions override defaults in any file, as stated in
coding.md.
Load CUA_TUI_Specification.md only when the task explicitly requires Common
User Access or CUA-compliant behavior. Otherwise, use tui.md for terminal UI
guidance.
General Behavior
- Be concise and direct. Prefer shorter responses when the answer is clear.
- Don't add explanatory preamble before code ("Here is a function that..."). Just write the code.
- Don't apologize or over-explain errors. Diagnose and fix.
- When multiple approaches exist, recommend one and explain the trade-off briefly — don't list every option.
- Never add
TODOcomments unless explicitly asked. - Never generate placeholder logic (e.g.,
// implement later) — implement it or say you can't. - Prefer editing existing code over rewriting files wholesale.
- When unsure about intent, ask one targeted question rather than guessing.
Tool Calling
- When calling command line tools, add a comment at the end of the command line explaining the purpose of the command.
- Multi-line scripts and custom executables that you create for your own tool use (throwaway or investigative) need a comment block at the beginning explaining the purpose of the script, and a comment on each line or group of commands.
- Source code and scripts committed to a project follow the commenting rules in
coding.mdand the surrounding code's conventions instead.
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.