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/puppyone-ai/puppyone/agents-mdgit clone --depth 1 https://github.com/puppyone-ai/puppyoneWhat 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 | $0.05311 | $0.05311 |
| Opus 5 | $0.02655 | $0.02655 |
| Sonnet 5 | $0.01062 | $0.01062 |
| Haiku 4.5 | $0.00531 | $0.00531 |
Grade A, and why
puppyone 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 — 415 lines — stays where its author put it; the contents beside it link to each section on GitHub.
PuppyOne (ContextBase)
AI assistants working on this codebase: the canonical Version Engine architecture is in
docs/architecture/01-version-engine.md. PuppyOne is Git-native at the version layer: stockgittalks tohttps://<host>/git/ap/<access_key>.git, while Web/API/puppyone fswrites converge through the Product Operation Adapter. Do not introduce the removed legacy wire protocol, external version package, or old source naming.
Overview
PuppyOne is a cloud file system built for AI Agents, centered around two core pillars: Connect and Collaborate.
It aggregates information scattered across various sources into a unified Context Space, while providing a complete infrastructure for multi-party collaboration between humans and agents — authentication, access control, version history, audit logging, and backup/rollback. Through the file system, bash, and the MCP protocol, any agent can read and write this ContextBase just like a local file system.
Connect
- Multi-source data connectors — OAuth connectors for 15+ platforms including Notion, GitHub, Gmail, Google Drive, Linear, Airtable, and more; also supports URL scraping, database connections, and custom scripts
- Bidirectional local folder sync — Real-time sync between local directories and the cloud Context Space via Git Remote and Puppyone CLI entry points
- MCP protocol exposure — Generates standard MCP interfaces for each agent or endpoint; any MCP-compatible client (Claude Desktop, Cursor, etc.) can connect directly
- Code sandbox — Securely execute code in isolated Docker/E2B containers; agents can invoke sandbox endpoints remotely
Collaborate
- Authentication & access control — JWT for human users + Access Key for machine authentication; agent-level node access permissions
- Version history & rollback — File-level version management, arbitrary version diff comparison, one-click rollback; folder-level snapshots
- Audit logging — Records all operations (who did what to which node, and when), fully traceable
- Collaborative editing — Checkout/commit workflow, locking mechanism, conflict detection and resolution
- Structured data management — Cloud file system (folders/JSON/Markdown/files), JSON Pointer table operations
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 · 415 lines · 5,311 tokens per session scan A f5d6b04122d6
puppyone AGENTS.md is an instructions file published in the GitHub repository puppyone-ai/puppyone (1,299 stars, last pushed 29d ago), licensed Apache-2.0. It adds 5,311 tokens to every session, about $0.0266 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
agent-library CLAUDE.md
Instructions for arcadeai-labs/agent-library, covering claude.md, project overview, key technologies, current status and development commands.
elephant-agent AGENTS.md
AGENTS.md instructions for agentic-in/elephant-agent, covering elephant agent entry, product north star, read first, non-negotiable rules and canonical commands.
elephant-agent copilot-instructions.md
Copilot instructions for agentic-in/elephant-agent, covering github copilot instructions, source of truth order, review priorities, harness-specific checks and validation expectations.
colin CLAUDE.md
Claude Code instructions for PrefectHQ/colin, covering colin development guidelines, required development workflow, documentation requirements, user-facing documentation (docs/) and docs/architecture.md.
colin AGENTS.md
AGENTS.md instructions for PrefectHQ/colin, a project described as: A context engine that treats skills as software.
mini-swe-agent copilot-instructions.md
Instructions for SWE-agent/mini-swe-agent, covering style guide, test style, bad and good.