agentleFS
Sign inStart free
taixuanntaixuann/development-space/AGENTS.md

development-space

AGENTS.md1 starsChanged 12 days agoMITView on GitHub

A software-development workshop for engineering skills, agent workflows, and local checkouts of active tool repositories. This repository is a thin workshop. It owns: It does not own scheduled automation, consumer-space data, or a nested tools/<space>/ tree. Product source of truth for active cloud work (for example raw-viewer) lives in the cloud environment / GitHub; clone into repos/ only when you need a local fix. Tools here use Python, Swift, or Rust. A tool-level AGENTS.md overrides the Python wording below. For…

The file

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

9 rules. About 0 are common, 1 is distinctive, and 8 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 8

    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:

    All agent operations in this workspace strictly follow the Agent Skills specification and official upstream skills from …

    line 23
  2. Very specific:

    Before writing code, use create-technical-spike for uncertain technical proofs and create-specification / …

    line 36
  3. Very specific:

    Use superpowers:subagent-driven-development to execute tasks in isolated contexts.

    line 42
  4. Very specific:

    When bugs occur, run superpowers:systematic-debugging to trace root causes before proposing fixes.

    line 44
  5. Very specific:

    Run tests inside the product repository (for example pytest under repos/<tool>/), not at the workshop root.

    line 45
  6. Very specific:

    Run agentic-eval to validate payloads against JSON schemas.

    line 49
  7. Very specific:

    Run docs-sync-audit to ensure README, CLI help text, and documentation match code changes with zero drift.

    line 50
  8. Very specific:

    Path Agnosticism: Never hardcode user home paths.

    line 54
  9. Distinctive:

    Git safety: Never force-push or merge unverified changes to main.

    line 58

≈ “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.

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

Development Space

A software-development workshop for engineering skills, agent workflows, and local checkouts of active tool repositories.

Overview

This repository is a thin workshop. It owns:

  • Shared engineering skills under .agents/skills/
  • Workshop policy and routing (AGENTS.md, docs/)
  • Optional local product checkouts under repos/<tool-name>/ (each an independent Git repository)

It does not own scheduled automation, consumer-space data, or a nested tools/<space>/ tree. Product source of truth for active cloud work (for example raw-viewer) lives in the cloud environment / GitHub; clone into repos/ only when you need a local fix.

Work Phases

Tools here use Python, Swift, or Rust. A tool-level AGENTS.md overrides the Python wording below.

For a platform or app, work in three phases: think and decide, lay the foundation, then build. During /brainstorming, read docs/AGENTS.md and propose the skills it lists. Run a proposed skill only after the user agrees.

Upstream Skills & Standards

All agent operations in this workspace strictly follow the Agent Skills specification and official upstream skills from github/awesome-copilot:

  • Package & CLI Architecture: python-pypi-package-builder (PyPA standard src/ layout, pyproject.toml, SemVer, uv tool packaging).
  • Modularity & Structure: folder-structure-blueprint-generator (dynamic Separation of Concerns) and code-exemplars-blueprint-generator (canonical style enforcement).
  • Execution Discipline: Superpowers methodology (superpowers:writing-plans, superpowers:subagent-driven-development, superpowers:test-driven-development, superpowers:systematic-debugging).
  • Stream & Batch Formatting: batch-script-progress (clean stdout for data/JSON, stderr for progress/diagnostics, standardized exit codes).
  • Quality & Security Gates: agentic-eval (self-critique schema validation), agent-owasp-compliance, mcp-security-audit, and docs-sync-audit (zero doc drift).
  • UI & Presentation: impeccable for product-grade webview/frontend design.
  • Local Persistence: sqlite-data-modeling (planned, not installed) for embedded databases and offline caches.

Development Workflow

1. Planning & Scaffolding

  • Before writing code, use create-technical-spike for uncertain technical proofs and create-specification / create-architectural-decision-record for architectural clarity.
  • Break implementation down into small, verifiable steps using superpowers:writing-plans.
  • Convert plan phases into GitHub tracking issues via create-github-issues-feature-from-implementation-plan.
  • Scaffold new tools using python-pypi-package-builder with standard pyproject.toml and clean src/ package layout.

2. Implementation & Testing

  • Use superpowers:subagent-driven-development to execute tasks in isolated contexts.
  • Strictly adhere to superpowers:test-driven-development: write failing tests first, implement the minimum code to pass, then refactor.
  • When bugs occur, run superpowers:systematic-debugging to trace root causes before proposing fixes.
  • Run tests inside the product repository (for example pytest under repos/<tool>/), not at the workshop root.

3. Verification & Delivery

  • Prior to completing a task, enforce superpowers:verification-before-completion with fresh execution logs.
  • Run agentic-eval to validate payloads against JSON schemas.
  • Run docs-sync-audit to ensure README, CLI help text, and documentation match code changes with zero drift.

Core Engineering Conventions

  1. Path Agnosticism: Never hardcode user home paths. Use pathlib.Path.home(), $HOME, or environment variables.
  2. Configuration Standards: Store user-specific configuration outside repositories in standard XDG paths (~/.config/<tool-name>/), using tracked *.example.yaml templates in source.
  3. CLI Design: Support standard flags (--help, -h, --json), non-interactive overrides (--yes/--force), and predictable exit codes (0 success, 1 runtime error, 2 validation error).
  4. Local checkouts: Active local checkouts live under repos/<tool-name>/. Each is its own Git repository and is ignored by the workshop root Git. Prefer cloning from the remote/cloud source of truth when starting local work.
  5. Git safety: Never force-push or merge unverified changes to main.

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.

Check your setup, free