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/intentdriven/abcd/agents-mdgit clone --depth 1 https://github.com/intentdriven/abcdWhat 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 | $0.03564 | $0.03564 |
| Opus 5 | $0.01782 | $0.01782 |
| Sonnet 5 | $0.00713 | $0.00713 |
| Haiku 4.5 | $0.00356 | $0.00356 |
Grade A, and why
abcd 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 yesterday.
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.
How it starts
The opening of the file, as written. The whole thing — 242 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
abcd rule loader
This repository uses the abcd modular rules loader. On UserPromptSubmit, a hook
recall-matches the prompt against keyword triggers declared in the plugin-bundled
default domains and <repo>/.abcd/rules.json, and injects only the matched
domain rules into context — instead of force-loading the full ruleset every turn.
A prompt that matches no domain injects nothing (zero added tokens).
- Inspect rules:
abcd rulesrenders the active set;abcd rules <DOMAIN>(case-insensitive) scopes to one domain. - Per-repo overrides: edit
<repo>/.abcd/rules.json. It is{"schema_version": 1, "disabled": false, "domains": {}}— add a domain key to override a default per-field (e.g.{"ROADMAP": {"state": "dormant"}}silences it while keeping its rules) or to declare a custom domain ({"recall": [...], "rules": [...]}). - Kill switch: set
"disabled": trueat the top of.abcd/rules.json. - Explicit activation: start a prompt with
*<DOMAIN>(e.g.*COMMITTING,*PII) to inject that domain unconditionally — overrides adormantstate, but never the kill switch.
Default domains
COMMITTING, DOCUMENTATION, ROADMAP, ISSUES, INTENTS, LIFEBOAT, PII,
OPINIONS. Each carries recall keywords and its rules, bundled in the abcd
binary; a repo overrides them per-field via .abcd/rules.json. OPINIONS
points at the canonical conventions under .abcd/development/principles/ rather
than copying them.
Reset triggers
SessionStart and PreCompact clear the per-session dedup ledger, so a matched
domain re-injects on the next prompt (the event-driven refresh that recovers
after compaction). Within a session the hook does not re-inject unchanged rules.
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.
- yesterday First seen · 242 lines · 3,564 tokens per session scan A 12bfab5b6bed
abcd AGENTS.md is an instructions file published in the GitHub repository intentdriven/abcd (3 stars, last pushed 2d ago), licensed MIT. It adds 3,564 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 2026-08-31.
Other instructions, from other repositories
nexent AGENTS.md
Instructions for ModelEngine-Group/nexent, covering agents, available skills, project overview, developer commands and backend (python 3.11).
turbo-flow CLAUDE.md
Claude Code instructions for marcuspat/turbo-flow, covering claude.md — turbo flow, what this repo is, key files and conventions.
belmont AGENTS.md
Instructions for blake-simpson/belmont, covering agents, self-learning loop — consult knowledge/knowledge.md first, both invocation paths or it's not done, parallel mode invariants — enforced by multiple layers and notes.
orc CLAUDE.md
Claude Code instructions for spencermarx/orc, covering claude.md, openspec instructions, migration changelog, what orc is and architecture.
orc AGENTS.md
AGENTS.md instructions for spencermarx/orc, covering openspec instructions, landing the plane (session completion) and open code review instructions.
belmont CLAUDE.md
Instructions for blake-simpson/belmont, a project described as: An agent agnostic, project agnostic, AI toolkit for running structured coding sessions with AI coding agents.