thinkhaven-method-kit
AGENTS.md0 starsChanged 4 months agoMITView on GitHub
Context for AI coding agents (Claude Code, Codex, Cursor, Gemini, and any other tool reading this file) working in the Decision Architecture repository. This file is the canonical agent context. CLAUDE.md redirects here for backward compatibility. Decision Architecture is a structured-thinking toolkit for making better decisions before you build. It ships as a Claude Code plugin (skills, agents, commands, prompts, frameworks). The prompts and frameworks are also designed to be portable to any LLM-driven workflow. Tools other than Claude Code…
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 repository15 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- ai-graduationSkill
- arbitrage-auditSkill
- assumption-mapperSkill
- b2b-idea-vettingSkill
- codifySkill
- decision-briefSkill
- design-thinking-liteSkill
- grill-meSkill
- grill-with-docsSkill
- innovation-diagnosticSkill
- kill-goSkill
- preflightSkill
- problem-framingSkill
- pull-frameworkSkill
The file
We split this AGENTS.md into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
No consulting jargon, no corporate speak
- Distinctive: line 47
Every skill includes the footer linking to ThinkHaven, BMAD-METHOD, and data-product-operator
≈ “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.
Files
AGENTS.md
Context for AI coding agents (Claude Code, Codex, Cursor, Gemini, and any other tool reading this file) working in the Decision Architecture repository.
This file is the canonical agent context. CLAUDE.md redirects here for backward compatibility.
Project
Decision Architecture is a structured-thinking toolkit for making better decisions before you build. It ships as a Claude Code plugin (skills, agents, commands, prompts, frameworks). The prompts and frameworks are also designed to be portable to any LLM-driven workflow.
Voice & Style
- Imperative, practitioner language: "Ask X", "Probe Y", "Surface Z"
- No consulting jargon, no corporate speak
- Direct, specific, actionable
- Challenge assumptions by default — anti-sycophancy is a feature
File Structure
| Path | Purpose |
skills/<name>/SKILL.md | Background-context skills (Claude Code: activate on keyword match) |
agents/<name>.agent.md | Foreground agents (Claude Code: invoke via slash command or @-mention) |
commands/<name>.md | User-facing entry points. Invoke by bare name (/decide, /codify); fully qualified form is /thinkhaven-method-kit:<name> (commands namespace by the plugin.json name — verified by live install 2026-07-05) |
prompts/ | Standalone prompts that work with any LLM |
frameworks/ | Reference documents for decision practice |
docs/brainstorms/ | Origin requirements documents |
docs/plans/ | Implementation plans |
.claude-plugin/plugin.json | Plugin manifest |
Tools other than Claude Code can read the prompts and frameworks directly. Skills and agents follow the Claude Code plugin format; equivalent invocation patterns vary by tool.
Consuming projects (not this repo) get a decisions/ folder of durable one-decision-per-file records, written by the codify skill at the end of the loop. This line documents the convention for contributors; the skill's own runtime check is what surfaces it to consumers.
Conventions
- Each skill: YAML frontmatter (
name,description,user-invocable) plus imperative guidance. Cross-reference related skills, don't duplicate content. - Each agent (added v1.1):
<name>.agent.mdinagents/. Frontmatter isname,description(with activation keywords),model,tools. Body is imperative system prompt with anti-sycophancy core embedded. Voice matches skills; no named persona characters. Agents are self-contained — they inline the question lists and templates they need rather than reading sibling skill files at runtime. - Plugin manifest at
.claude-plugin/plugin.jsondeclares metadata. Directory-based discovery handlesskills/,commands/,agents/.
Quality Bar
- Useful output on first use (no learning curve)
- Works for coding AND non-coding decisions
- Session takes under 15 minutes
- Output is shareable markdown
- Every skill includes the footer linking to ThinkHaven, BMAD-METHOD, and data-product-operator
Framework Sources
- Decision architecture wiki (curated practitioner notes)
- BMAD-METHOD analyst (Mary) and architect (Winston) patterns
- BMAD Creative Intelligence Suite (Victor, Dr. Quinn, Maya)
- Nate Jones, Matt Pocock, Stuart Winter-Tear, Alex MH Smith
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.