Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/dobermancore/doberman-core/agents-mdgit clone --depth 1 https://github.com/DobermanCore/Doberman-CoreWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/dobermancore/doberman-core/agents-md)<a href="https://agentmods.dev/instructions/dobermancore/doberman-core/agents-md"><img src="https://agentmods.dev/badge/instructions/dobermancore/doberman-core/agents-md.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.03567 | $0.03567 |
| Opus 5 | $0.01784 | $0.01784 |
| Sonnet 5 | $0.00713 | $0.00713 |
| Haiku 4.5 | $0.00357 | $0.00357 |
Grade A, and why
Doberman-Core AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
Copies of this mod
1 near-identical copy found in the catalogue:
- Doberman-Core CLAUDE.md — 86% identical, 144 lines differ
How it starts
The opening of the file, as written. The whole thing — 252 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md — Doberman Operating Manual
This is the init file for this repository. It governs any AI agent (Claude Code, Codex, Cursor, etc.) working here. It is the HOW. The WHAT — the feature roadmap and per-slice detail — lives in
README.md(public summary) and, if present in your working tree, the project development plan you keep alongside it. Content is identical toAGENTS.md; keep the two in sync (or symlink one to the other). If this file conflicts with a casual prompt instruction, this file wins unless a human overrides it in writing.
0. On startup (every session)
- Read this entire file, then skim
README.mdfor the current feature set, versioning, and roadmap. If you keep a local development plan alongside the repo, read the relevant slice there for the extra detail (objective, files, edge cases, suggested tests, suggested commit message). git status+git log --oneline -5.- Pick the next slice (the lowest-numbered unmerged slice in the planned order).
- If the repo lacks scaffolding (
pyproject.toml/.github/workflows/ci.yml), do Slice 0 (Bootstrap) first (§6). - Execute exactly one slice via The Slice Loop (§5), then STOP and report. Stop at every review checkpoint.
1. What this repository is
Doberman is an adaptive authorization layer for coding agents, released publicly under Apache-2.0.
It distributes as the doberman package (src/doberman/) and is designed to be genuinely functional on its own —
schema, interfaces, the runtime harness, adapters, and the built-in rules all live here. It is not crippleware.
Doberman is built to be extensible: it declares stable interfaces (Rule / Detector / AuthProvider /
AuditSink) and a runtime registry that discovers implementations via Python entry points. Additional packages can
register their own rules, detectors, auth providers, or audit sinks without Doberman importing them by name — the
core never takes a static dependency on any plugin. With only doberman installed, it works (built-in protection);
a plugin package is imported only after you name it with doberman plugins enable <entry-point name> — nothing
loads because it is merely installed.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today Changed · +1 lines · +20 tokens per session e934bb2a6e92
- 4d ago First seen · 251 lines · 3,547 tokens per session scan A 09725b842087
Doberman-Core AGENTS.md is an instructions file published in the GitHub repository DobermanCore/Doberman-Core (197 stars, last pushed today), licensed Apache-2.0. It adds 3,567 tokens to every session, about $0.0178 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
llm-safe-sql AGENTS.md
AGENTS.md instructions for hyuga611/llm-safe-sql, covering notes for coding agents, what this project is, rules that are not style preferences, where things are and before you say you are done.
bifrost AGENTS.md
AGENTS.md instructions for maximhq/bifrost, covering agents.md — bifrost ai gateway, what is bifrost?, repository layout, go workspace and build, test & dev commands.
caracal file-header.instructions.md
Use when creating or editing any source file. Enforces the mandatory copyright header format for all files.
caracal console-example-naming.instructions.md
Use when writing, editing, or reviewing web console help, info pages, field examples, guided setup copy, or example-facing web console UI text. Enforces the approved Caracal demo naming universe.
caracal code-style.instructions.md
Use when writing, editing, or reviewing any code. Enforces naming discipline, comment standards, and variable hygiene across the entire codebase.
caracal directory-instructions.instructions.md
Use when creating, updating, or reviewing any instructions.md file in the codebase.