Spark-To-Paper-SkillsSpark-To-Paper-Skills/paperjury-codex/codex/AGENTS.md

paperjury-codex / codex

AGENTS.md222 starsChanged 5 months agoMITView on GitHub

PaperJury Codex is a Codex-first skill. Use codex/phase-contracts.md as the prompt/schema/isolation contract and follow codex/runtime.md for runtime orchestration. All paths in this file are repository-root relative. For any Codex task that uses, ports, audits, or changes PaperJury: 1. Read SKILL.md. 2. If the task involves review, auto mode, fan-out, or subagents, read codex/runtime.md. 3. For Codex semantic phase work, read codex/phase-contracts.md. 4. For engine changes, read references/review-engine-v3.md and references/ledger-schema.md. 5. For user-facing install or usage docs, also read codex/AGENT-GUIDE.md. -…

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 31,774 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 repository2 other files it gives its agents.
- AGENTS.mdAGENTS.md
- paperjurySkill

The file

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

11 rules. About 0 are common, 8 are distinctive, and 3 are very specific to this project.

≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.

Common
Distinctive
Very specific

Distinctive:
All paths in this file are repository-root relative.
line 7
Very specific:
If the task involves review, auto mode, fan-out, or subagents, read codex/runtime.md.
line 14
Distinctive:
Do not require Claude Workflow files for Codex operation.
line 23
Very specific:
Use codex/phase-contracts.md for semantic phase inputs, outputs, isolation lines, and deterministic validation rules.
line 24
Very specific:
Use node scripts/*.js for deterministic guards.
line 26
Distinctive:
Use Codex subagents for semantic fan-out only when the user has explicitly authorized multi-agent work.
line 27
Distinctive:
Keep reviewer and juror prompts isolated: inline only the manuscript/context they may judge, pass no project file paths, no ledger, no …
line 30
Distinctive:
Never edit a manuscript without explicit author sign-off.
line 33
Distinctive:
Do not hardcode local paper paths into the skill.
line 38
Distinctive:
Before finishing a migration or runtime change, run static checks when possible:
line 42
Distinctive:
Do not run a real paper-review scenario unless the user explicitly asks for it.
line 48

≈ “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

PaperJury Codex Instructions

PaperJury Codex is a Codex-first skill. Use codex/phase-contracts.md as the prompt/schema/isolation contract and follow codex/runtime.md for runtime orchestration.

All paths in this file are repository-root relative.

Load Order

For any Codex task that uses, ports, audits, or changes PaperJury:

Read SKILL.md.
If the task involves review, auto mode, fan-out, or subagents, read codex/runtime.md.
For Codex semantic phase work, read codex/phase-contracts.md.
For engine changes, read references/review-engine-v3.md and references/ledger-schema.md.
For user-facing install or usage docs, also read codex/AGENT-GUIDE.md.

Codex Runtime Rules

Do not require Claude Workflow files for Codex operation.
Use codex/phase-contracts.md for semantic phase inputs, outputs, isolation lines, and deterministic validation rules.
Use node scripts/*.js for deterministic guards.
Use Codex subagents for semantic fan-out only when the user has explicitly authorized multi-agent work. If subagents are unavailable or unauthorized, run a clearly labeled degraded path and do not claim full isolation.
Keep reviewer and juror prompts isolated: inline only the manuscript/context they may judge, pass no project file paths, no ledger, no prior-round reports, and no peer outputs.
Never edit a manuscript without explicit author sign-off. Auto mode requires up-front policy sign-off and queues anything outside the authorized envelope.
Store mutable state in the project-owned .paper-review/ directory: LEDGER.json, rendered LEDGER.md, journal.jsonl, spine.json, and any stable CONVENTIONS.md.
Do not hardcode local paper paths into the skill.

Validation

Before finishing a migration or runtime change, run static checks when possible:

node scripts/doctor.js --project <paper-project>
python <codex-skill-creator>/scripts/quick_validate.py <paperjury-dir>
git status --short

Do not run a real paper-review scenario unless the user explicitly asks for it.

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