agentleFS
tailscale-devtailscale-dev/ScaleTail/AGENTS.md

ScaleTail

AGENTS.md2.2k starsChanged 19 days agoMITView on GitHub

ScaleTail is a collection of Docker Compose stacks. Each directory in services/ is one stack, which usually runs a self-hosted application behind a Tailscale sidecar container. From each changed service directory, validate Compose. The command gives each required variable a dummy value, as CI does, so you do not edit .env: It must exit with status 0. CI fails on "variable is not set" warnings. From the repository root, lint all Markdown with the rumdl version that .github/workflows/linting.yml pins in…

  • Reads credentials

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

7 rules. About 0 are common, 2 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  • could be automatic: a setting could enforce it
  1. Very specific:

    Read the service's compose.yaml, .env, and README, including "Deviations from the standard setup".

    line 14
  2. Distinctive:

    Keep changes focused on the requested service or documentation.

    line 17
  3. Very specific:

    Add, rename, or remove a service: the services/<name>/ directory and its row in the root README.md.

    line 22
  4. Very specific:

    Every .env file is tracked and published.

    line 28
  5. Very specific:

    Start test stacks from a copy outside the repository, such as cp -R services/<name> /tmp/<name>, and put real values only in the copy.

    line 30
  6. Distinctive:

    Before you commit, run git status and stage only the files you meant to change.

    line 31
  7. Very specific:

    When you have an auth key, start the copied stack and test its main function through the Tailnet.

    line 55

≈ “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

Agent guidance

ScaleTail is a collection of Docker Compose stacks. Each directory in services/ is one stack, which usually runs a self-hosted application behind a Tailscale sidecar container.

Where the rules live

  • CONTRIBUTING.md: how to add, update, verify, and submit a service. Follow it.
  • templates/service-template/: the starting point for every service. Its comments explain each setting.
  • documentation/standard-setup.md: the shared setup. A service README lists only what differs, under "Deviations from the standard setup".
  • .github/workflows/: the Compose validation and Markdown lint that CI runs.

Before you edit

  • Read the service's compose.yaml, .env, and README, including "Deviations from the standard setup".
  • Confirm ports, commands, volumes, variables, and health checks in the image's upstream docs.
  • Preserve documented exceptions. Some stacks use other Compose service keys or container names, and other containers or .env values refer to those names.
  • Keep changes focused on the requested service or documentation.
  • Apply the rules to new stacks and to the parts of an existing stack that you change. Existing stacks are brought in line over time, so do not fix the rest of a stack unless that is the task.

Files that change together

  • Add, rename, or remove a service: the services/<name>/ directory and its row in the root README.md. Insert a new row in alphabetical order by the bold name, and leave the other rows alone.
  • Make existing users act, such as with a new required secret, a moved volume, or a changed SERVICE: add an "Upgrading" section to the service README. SERVICE also names the Tailnet device, the containers, and most data folders.
  • Change the template: also update documentation/standard-setup.md and CONTRIBUTING.md where they describe the same setting. Validate it from templates/service-template/ with SERVICE=dummy IMAGE_URL=dummy docker compose config --quiet, because the template leaves both values empty.

Secrets and local data

  • Every .env file is tracked and published. Never commit an auth key, password, token, or other credential.
  • Leave each secret empty in .env. When the stack cannot start without it, read it in compose.yaml as ${VAR:?Set VAR in .env}, and add a # Required: comment above it in .env that says what to set, such as how to generate it. Keep TS_AUTHKEY optional, because only the first start needs it.
  • Start test stacks from a copy outside the repository, such as cp -R services/<name> /tmp/<name>, and put real values only in the copy. A running stack writes ./config, ./ts/state (with the device key), and its data folders next to compose.yaml.
  • Before you commit, run git status and stage only the files you meant to change.

Verification

From each changed service directory, validate Compose. The command gives each required variable a dummy value, as CI does, so you do not edit .env:

env $(grep -ohE '\$\{[A-Za-z0-9_]+:\?' compose.yaml .env | sed -E 's/^\$\{//; s/:\?$/=dummy/' | sort -u) docker compose config --quiet

It must exit with status 0. CI fails on "variable is not set" warnings.

From the repository root, lint all Markdown with the rumdl version that .github/workflows/linting.yml pins in its version: input, and check your changes for whitespace errors:

version=$(sed -nE 's/.*version: "([0-9.]+)".*/\1/p' .github/workflows/linting.yml)
uvx "rumdl@$version" check --config .markdownlint.yml .
git diff --check origin/main

Without uv, download that version from the rumdl releases and run rumdl check --config .markdownlint.yml ..

Lint the whole tree, because renaming a service breaks a link in the root README.md. git diff --check origin/main also covers staged and committed changes, but not untracked files, so stage new files first. In a fork, compare with the main branch of this repository instead, for example upstream/main.

When you have an auth key, start the copied stack and test its main function through the Tailnet. Run each health check command in its container with docker exec.

Report the checks you ran. State any checks that credentials, hardware, permissions, or unavailable tools prevented.

Pull requests

  • Title the pull request Service: what changes, such as Umami: new service. For several services, list them or use All services:. For other areas, use CI:, Docs:, or Template:.
  • Fill in the pull request template. Maintainers apply the labels.

Download now (ZIP)

Trending

Less talk lately.thin dataabout the whole repository

Dec 2025: 0 mentionsJan 2026: 1 mentionFeb 2026: 0 mentionsMar 2026: 0 mentionsApr 2026: 2 mentionsMay 2026: 1 mentionJun 2026: 0 mentionsJul 2026: 2 mentionsAug 2026: 0 mentionsSep 2026: 0 mentionsOct 2026: 1 mentionNov 2026: 0 mentions
Dec 2025Jun 2026Nov 2026
Mentions a month of the tailscale-dev/ScaleTail repository that our weekly crawl found: only 7 in all, so this is a hint, not a trend. There is no adoption data for this file yet.

Who's talking about tailscale-dev/ScaleTail

The biggest 5 of 7 mentions linking to the repository that our crawl found. Each line is their own headline.

See all 7

Talked about on GitHub and Bluesky (7 times).

On GitHub · 4

On Bluesky · 3

  • Other versions1 project uses a version of this.

    Word for word · 1

    1. 2Tiny2Scale/ScaleTail★ 2.2k100% same

    Edited · 0

    None found.

    Wrote their own version · 0

    None found.

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • 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.

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.