NITISH-R-GNITISH-R-G/hackerrank-orchestrate-skills/AGENTS.md

hackerrank-orchestrate-skills

AGENTS.md9 starsChanged 22 days agoMITView on GitHub

This repository holds Orchestrate skills: plain-markdown playbooks for the HackerRank Orchestrate hackathon, plus a Python toolkit that audits submissions. Any agent can use them: Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Aider, or a plain chat model with the files pasted in. Skills are ordinary markdown. The SKILL.md frontmatter is only used by Claude Code's skill loader; every other agent reads the body text. Phase order: orchestrate -> orchestrate-build -> orchestrate-test-fix -> orchestrate-submit. orchestrate-interview runs in parallel from…

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 31,542 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 repository12 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- .cursor/rules/orchestrate.mdcCursor rule
- llms.txtllms.txt
- orchestrate-buildSkill
- orchestrate-interviewSkill
- orchestrate-packageSkill
- orchestrate-referenceSkill
- orchestrate-scoreSkill
- orchestrateSkill
- orchestrate-submitSkill
- orchestrate-test-fixSkill

The file

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

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

≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
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

Very specific:
Read skills/INDEX.md. It maps each situation to one skill.
line 9
Very specific:
Start with skills/orchestrate/SKILL.md.
line 10
Distinctive:
Never tune against sample rows until one number matches.
line 18
Distinctive:
Keep the output contract exact: column names, allowed values, separators, row count.
line 20
Very specific:
Never write a number you did not measure.
line 21
Distinctive:
Do not overwrite a starter CLAUDE.md or AGENTS.md in the challenge repo.
line 22
Distinctive:
Keep a safe fallback output banked before any experiment.
line 23
Distinctive:
Before changing anything, state the metric it should move and the rows it could break.
line 24
Very specific:
Keep skills in skills/<name>/SKILL.md with name and description frontmatter.
line 30
Distinctive:
Do not add a new skill when an existing one covers the job.
line 31

≈ “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 - instructions for any AI coding agent

This repository holds Orchestrate skills: plain-markdown playbooks for the HackerRank Orchestrate hackathon, plus a Python toolkit that audits submissions.

Any agent can use them: Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Aider, or a plain chat model with the files pasted in. Skills are ordinary markdown. The SKILL.md frontmatter is only used by Claude Code's skill loader; every other agent reads the body text.

How to use the skills (any agent)

Read skills/INDEX.md. It maps each situation to one skill.
Start with skills/orchestrate/SKILL.md. It holds the phase order and the failure routing.
Go to the phase skill it names. Each skill lists its references/ files. Read only the ones you need.
Re-open orchestrate whenever a step fails.

Phase order: orchestrate -> orchestrate-build -> orchestrate-test-fix -> orchestrate-submit. orchestrate-interview runs in parallel from mid-build. orchestrate-reference is lookup only.

Rules for the agent working on a submission

Never tune against sample rows until one number matches. That is overfitting.
Model describes, code decides. Decisions use booleans and enums, never free text.
Keep the output contract exact: column names, allowed values, separators, row count.
Never write a number you did not measure. Use the provenance labels in skills/orchestrate-interview/references/CONSTANTS-template.md.
Do not overwrite a starter CLAUDE.md or AGENTS.md in the challenge repo. Append to it.
Keep a safe fallback output banked before any experiment.
Before changing anything, state the metric it should move and the rows it could break.

Rules for working on this repository

Python 3.10+. Run tests with python -m pytest -q.
Tests must pass on Windows and Linux. Full git history is required for commit-provenance tests (git fetch --unshallow on a shallow clone).
Keep skills in skills/<name>/SKILL.md with name and description frontmatter. Keep each skill's detail in references/.
Do not add a new skill when an existing one covers the job. Edit the existing skill instead.
Historical documents live in docs/history/. Do not link to them from skills.

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