agentleFS
stormlightlabsstormlightlabs/thunderus/CLAUDE.md

thunderus

CLAUDE.md4 starsChanged 57 days agoApache-2.0View on GitHub

thndrs is a minimal coding agent. Its default behavior is an explicit prompt, project instructions, local tools, and durable sessions. Use the writing-docs skill for every piece of prose you produce here. That means documentation, commit messages, pull request bodies, issue bodies, review comments, and the replies you write in chat. Load it before you write, not after, and read references/tells.md when the prose is longer than a paragraph. Nothing checks this and nothing will: style is a judgement a…

The file

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

6 rules. About 0 are common, 4 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    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
  1. Very specific:

    Use the writing-docs skill for every piece of prose you produce here.

    line 8
  2. Very specific:

    Use //! module docs and /// docs for exported or important symbols.

    line 49
  3. Distinctive:

    Use enums, structs, and newtypes for meaningful values.

    line 51
  4. Distinctive:

    Keep parsing, policy, rendering projection, and validation pure where practical.

    line 52
  5. Distinctive:

    Prefer concrete helpers to traits.

    line 57
  6. Distinctive:

    Ensure documentation stays up to date as feature work is completed.

    line 74

≈ “Common” is an estimate: at least 1 in 5 of the 27,641 CLAUDE.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
CLAUDE.md

thndrs

thndrs is a minimal coding agent. Its default behavior is an explicit prompt, project instructions, local tools, and durable sessions.

Writing

Use the writing-docs skill for every piece of prose you produce here. That means documentation, commit messages, pull request bodies, issue bodies, review comments, and the replies you write in chat. Load it before you write, not after, and read references/tells.md when the prose is longer than a paragraph.

Nothing checks this and nothing will: style is a judgement a script cannot make, and a check that cannot be certain would only teach you to route around it. The rule is blunt instead. Prose written without the skill reads like it, and the tells it catalogues have each been shipped here at least once.

Development

The working tree is user-owned. Treat Git as read-only unless a user requests a Git operation. The user may change Git state while agents work. Track files changed by your own actions so do not use Git status as the sole record of agent work.

Workspace

  • thndrs-agent: provider-neutral agent loop, contracts, and context control.
  • thndrs: CLI/TUI application and ACP server mode.

thndrs-agent is a reusable leaf library.

The thndrs application composes it with application adapters own filesystem discovery, session persistence, terminal I/O, and transport, which stay in application adapters.

Provider wire payloads do not appear in public library APIs.

Code Style & Quality

Module order:

  1. constants
  2. traits
  3. enums and impls
  4. structs and impls
  5. exported functions
  6. private functions
  7. tests.

Use //! module docs and /// docs for exported or important symbols.

  • Use enums, structs, and newtypes for meaningful values.
  • Keep parsing, policy, rendering projection, and validation pure where practical. Isolate filesystem, network, process, environment, and terminal effects.
  • Return Result for recoverable failures. Production unwrap, expect, panic!, todo!, and unimplemented! need a documented invariant.
  • Prefer concrete helpers to traits. Traits mark real boundaries such as storage, process execution, providers, clocks, or tool executors.
  • Tests cover behavior, errors, state transitions, serialization, and side effects with deterministic fakes where possible.

Checks

For Rust changes, run the narrowest relevant test, then:

cargo fmt
cargo clippy --workspace --fix --allow-dirty --allow-staged
cargo clippy --workspace
cargo test --workspace

Documentation in the doc site changes require pnpm --dir docs build. Ensure documentation stays up to date as feature work is completed.

Keep docs/src/content/docs/docs/internals and docs/src/content/docs/docs/development current when changes affect system architecture, runtime behavior, or contributor workflows.

Interactive TUI QA

Hands-on terminal checks run in a dedicated tmux session. See Interactive TUI QA for the session setup, input and resize commands, and teardown.

Root level and internal/ files do not require a docs build.

Workflow

thunderstorm is the development loop for this repository. An issue and its sub-issues are the work; a run is one pass over one of them, and a milestone groups several without being an issue at all. See internal/thunderstorm.md for statuses, review passes, and branch rules, and internal/guides/using-thunderstorm.md for the operator's view of the loop. Skills and commands live in .claude/; .agents/skills is a symlink to .claude/skills for harnesses that follow the AGENTS.md convention. AGENTS.md is a symlink to this file.

Download now (ZIP)

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.