agentleFS
Sign inStart free
ntaffziintaffzii/Skill-Agents/AGENTS.md

Skill-Agents

AGENTS.md4 starsChanged 2 months agoMITView on GitHub

Project-wide rules for the Antigravity agent. Global Scope Note: These default mode settings apply strictly to complex projects. If the current workspace is a simple script or scratchpad, disregard the deep-planning requirement and default to fast execution. - New feature / multi-file change / architecture decision: Always use Planning Mode. Apply the deep-planning skill before writing any code. - Bug fix with unknown root cause: Planning Mode. Apply deep-planning. - Bug fix that is an obvious one-line typo or off-by-one:…

The file

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

8 rules. About 0 are common, 4 are distinctive, and 4 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.

  • 4

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    New feature / multi-file change / architecture decision: Always use Planning Mode. Apply the deep-planning skill before writing any code.

    line 9
  2. Distinctive:

    When in doubt, default to Planning Mode.

    line 17
  3. Distinctive:

    If an Implementation Plan artifact already exists for the current thread/feature,

    line 26
  4. Distinctive:

    Update the existing artifact in place.

    line 28
  5. Distinctive:

    When a plan changes materially, add a one-line "Revision note" at the top of the artifact explaining what changed and why, so the human can …

    line 29
  6. Very specific:

    Never invents scope beyond what architect approved.

    line 41
  7. Very specific:

    Default to architect first for anything matching the Planning Mode rules above.

    line 43
  8. Very specific:

    Only hand off to implementer once a plan has been explicitly approved.

    line 44

≈ “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

AGENTS.md

Project-wide rules for the Antigravity agent.

Global Scope Note: These default mode settings apply strictly to complex projects. If the current workspace is a simple script or scratchpad, disregard the deep-planning requirement and default to fast execution.

Default Mode Settings

  • New feature / multi-file change / architecture decision: Always use Planning Mode. Apply the deep-planning skill before writing any code.
  • Bug fix with unknown root cause: Planning Mode. Apply deep-planning.
  • Bug fix that is an obvious one-line typo or off-by-one: Fast Mode is fine.
  • Refactor touching more than one file: Planning Mode.
  • Refactor within a single file, no behavior change: Fast Mode is fine.
  • Test generation / documentation / formatting: Fast Mode.

When in doubt, default to Planning Mode. A wasted plan review costs seconds; an unreviewed multi-file change that goes wrong costs much more.

Plan continuity (important)

This project has had recurring issues with the agent losing earlier plan context when the conversation gets long or the user changes direction mid-thread.

Rules:

  1. If an Implementation Plan artifact already exists for the current thread/feature, treat new chat instructions as edits to that plan, not as a new task. Update the existing artifact in place. Never silently discard prior plan content.
  2. When a plan changes materially, add a one-line "Revision note" at the top of the artifact explaining what changed and why, so the human can track drift over time.
  3. For any task expected to span multiple sessions, write the approved plan to PLAN.md at the project root as the source of truth, not just the in-conversation artifact. Read PLAN.md at the start of every session before proposing new work.

Model routing

This project uses two custom agents (see .agents/agents/):

  • architect — planning, spec-writing, and plan revisions. Runs on a stronger reasoning model. Never edits files directly.
  • implementer — executes approved plans. Runs on a faster/cheaper model. Never invents scope beyond what architect approved.

Default to architect first for anything matching the Planning Mode rules above. Only hand off to implementer once a plan has been explicitly approved.

Skills

  • Apply deep-planning automatically whenever a task matches the Planning Mode criteria above, even if the user does not explicitly ask for a plan.

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