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/hmans/beans/claude-mdgit clone --depth 1 https://github.com/hmans/beansWhat 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 | $0.01364 | $0.01364 |
| Opus 5 | $0.00682 | $0.00682 |
| Sonnet 5 | $0.00273 | $0.00273 |
| Haiku 4.5 | $0.00136 | $0.00136 |
Grade A, and why
beans CLAUDE.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 2d ago.
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 — 91 lines — stays where its author put it; the contents beside it link to each section on GitHub.
What we're building
Beans is an agentic-first issue tracker. Issues ("beans") live as markdown files in a .beans/ directory inside a project repo. The system has three interfaces:
- CLI (
beansbinary) — create, list, update, and query beans from the terminal - Terminal TUI — Bubbletea-based interactive interface
- Beans UI (
beans serve) — SvelteKit SPA served by an embedded Go HTTP server, communicating via GraphQL (queries, mutations, subscriptions over WebSocket)
The Beans UI is the primary development focus. It includes a backlog board, per-bean agent chat (spawning Claude Code processes), git worktree management, file change diffs, and terminal sessions.
Commits
- Use conventional commit messages ("feat", "fix", "chore", etc.) when making commits.
- Include the relevant bean ID(s) in the commit message (please follow conventional commit conventions, e.g.
Refs: bean-xxxx). - Mark commits as "breaking" using the
!notation when applicable (e.g.,feat!: ...). - When making commits, provide a meaningful commit message. The description should be a concise bullet point list of changes made.
Pull Requests
- When we're working in a PR branch, make separate commits, and update the PR description to reflect the changes made.
- Include the relevant bean ID(s) in the PR title (please follow conventional commit conventions, e.g.
Refs: bean-xxxx).
Project Structure
Key packages:
pkg/bean/— Bean data model, parsing, sorting, validation (no I/O)pkg/beancore/— Core engine: disk I/O, file watching, search indexing, worktree watchinginternal/graph/— GraphQL schema and resolvers (the API layer for both Beans UI and CLI)internal/agent/— Agent session manager: spawns Claude Code processes, parses JSONL output, pub/sub for real-time updatesinternal/worktree/— Git worktree lifecycle managementinternal/terminal/— PTY session management for embedded terminalsinternal/commands/— CLI command implementations (Cobra)frontend/— SvelteKit SPA (embedded into the Go binary via//go:embed)
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.
- 2d ago First seen · 91 lines · 1,364 tokens per session scan A 35d85f148d22
beans CLAUDE.md is an instructions file published in the GitHub repository hmans/beans (911 stars, last pushed 4mo ago), licensed Apache-2.0. It adds 1,364 tokens to every session, about $0.0068 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
Backlog.md AGENTS.md
AGENTS.md instructions for MrLesk/Backlog.md, covering read this first, simplicity-first implementation rules, ui descriptions, commands and development.
Backlog.md copilot-instructions.md
Copilot instructions for MrLesk/Backlog.md, covering ⚠️ important, commands, development, testing and configuration management.
Backlog.md CLAUDE.md
Claude Code instructions for MrLesk/Backlog.md, a project described as: Backlog.md - A tool for managing project collaboration between humans and AI Agents in a git ecosystem.
abap_wiki AGENTS.md
Instructions for Gixsy95/abap_wiki, covering agents.md - abapwiki operating contract, 1. identity and mission, 2. language, 3. architecture: the two planes and 4. inviolable rules.
lineark CLAUDE.md
Instructions for flipbit03/lineark, covering claude.md — lineark, what is this, workspace layout, key commands and writing online tests: the three pillars.
apex-accelerator agent-operating-frame.instructions.md
Shared operating frame for main step agents — read SKILL.md once, use apex-recall for cached lookups, never edit upstream artifacts. Pairs with each agent's body-level Operating frame H2.