K-Ops
AGENTS.md1 starsChanged 7 months agoMITView on GitHub
This repo is designed for agentic CLI workflows. Canonical rules: see OPERATING_RULES.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 32,671 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 repository9 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .claude/CLAUDE.mdCLAUDE.md
- CLAUDE.mdCLAUDE.md
- compile-wikiSkill
- ingest-sourcesSkill
- lint-healSkill
- qa-agentSkill
- render-outputSkill
- research-workflowSkill
The file
We split this AGENTS.md into the rules it gives an agent
9 rules. About 0 are common, 3 are distinctive, and 6 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
Distinctive · estimate
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 36
Prefer modifying a small number of files with high signal.
- Distinctive: line 37
Prefer precise edits over broad rewrites.
- Distinctive: line 39
When uncertain, mark uncertainty explicitly.
- Very specific: line 42
Use uv run kops install-agent-assets to sync canonical skills/templates to the Codex runtime directory.
- Very specific: line 43
Use uv run kops validate to confirm config loads before running workflows.
- Very specific: line 44
Prefer uv run kops compile --agent codex when you want Codex to do the compilation pass.
- Very specific: line 45
Use uv run kops extract-claims, extract-contradictions, and scorecard to keep the machine-readable quality layer current.
- Very specific: line 46
After extract-claims, run uv run kops verify-spans to confirm claim quote anchors exist in their sources, and uv run kops review-queue to …
- Very specific: line 47
Use uv run kops next-action for the single next repair + convergence verdict, and uv run kops consequence-gate --tier <...> before backing …
≈ “Common” is an estimate: at least 1 in 5 of the 41,581 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.
Files
AGENTS.md
This repo is designed for agentic CLI workflows.
Canonical rules: see OPERATING_RULES.md.
Roles
Source Ingestor
Goal: transform new raw sources into normalized source summaries.
Wiki Compiler
Goal: merge source summaries into durable concept pages and indexes.
Lint + Heal
Goal: detect contradictions, weak structure, missing backlinks, and unsupported claims.
Q&A Agent
Goal: answer using the vault, then file durable insights back into the vault.
Render Agent
Goal: convert the current vault into memos, briefings, slide outlines, diagrams, and plot specs.
Repo Contract
data/raw/holds source evidencenotes/Sources/holds per-source summariesnotes/Concepts/holds durable knowledge pagesnotes/Answers/holds generated answer memosnotes/Home.mdis the Obsidian entry pointnotes/TODO.mdtracks gaps and healing tasksnotes/Runbooks/Agent_Workflow_Quick_Reference.mdis the compact operator map for cross-CLI workflowsresearch/holds resumable research-run artifacts and stays separate from curated vault noteskops bootstrap --target <dir>creates a new blank starter vault with the same file structure and tooling
Behavioral Guardrails
- Prefer modifying a small number of files with high signal.
- Prefer precise edits over broad rewrites.
- Preserve provenance from source summaries into concept pages.
- When uncertain, mark uncertainty explicitly.
Codex-Specific Notes
- Use
uv run kops install-agent-assetsto sync canonical skills/templates to the Codex runtime directory. - Use
uv run kops validateto confirm config loads before running workflows. - Prefer
uv run kops compile --agent codexwhen you want Codex to do the compilation pass. - Use
uv run kops extract-claims,extract-contradictions, andscorecardto keep the machine-readable quality layer current. - After
extract-claims, runuv run kops verify-spansto confirm claim quote anchors exist in their sources, anduv run kops review-queueto surface everything awaiting human review. - Use
uv run kops next-actionfor the single next repair + convergence verdict, anduv run kops consequence-gate --tier <...>before backing a high-stakes answer or render with the vault's claims.
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
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.

