Borrowing it
Nothing to install: this file belongs to bayeslabs-rsi/Svatah. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/bayeslabs-rsi/Svatah/main/CLAUDE.mdgit clone --depth 1 https://github.com/bayeslabs-rsi/SvatahWrote 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/bayeslabs-rsi/svatah/claude-md)<a href="https://agentmods.dev/instructions/bayeslabs-rsi/svatah/claude-md"><img src="https://agentmods.dev/badge/instructions/bayeslabs-rsi/svatah/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/instructions/bayeslabs-rsi/svatah/claude-md"><img src="https://agentmods.dev/badge/instructions/bayeslabs-rsi/svatah/claude-md.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.02351 | $0.02351 |
| Opus 5 | $0.01175 | $0.01175 |
| Sonnet 5 | $0.00470 | $0.00470 |
| Haiku 4.5 | $0.00235 | $0.00235 |
Grade A, and why
Svatah 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 11d 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 — 167 lines — stays where its author put it; the contents beside it link to each section on GitHub.
svatah
Autoresearch plugin for Codex and Claude Code — structured experiment-driven code optimization with git worktrees, traces, and a local dashboard.
Architecture Overview
svatah/
├── plugins/sva/ # Main Python CLI + plugin package (svatah-cli v0.5.0)
│ ├── src/sva/
│ │ ├── cli.py # CLI entry point (196KB, ~5000 lines) — all `sva` subcommands
│ │ ├── core.py # Workspace init, graph ops, experiment lifecycle, git worktree mgmt
│ │ ├── dashboard.py # Flask dashboard (auto-started on init, port 8080+)
│ │ ├── dispatch.py # Explorer-cache for parallel subagent dispatch (claude-code only)
│ │ ├── frontier_strategies.py # Node ranking: argmax, top_k, epsilon_greedy, ucb, etc.
│ │ ├── experiment_db.py # SQLite index (rebuildable from graph.json — graph is authoritative)
│ │ ├── analyzer.py # Post-run analysis report generation
│ │ ├── config_parser.py # sva-config.md parsing and validation
│ │ ├── token_tracker.py # Token usage tracking per experiment/branch
│ │ ├── scratchpad.py # Bounded state summary for agent context windows
│ │ ├── locking.py # Advisory file locking (portalocker)
│ │ ├── version_check.py # CLI ↔ plugin manifest version consistency
│ │ ├── literature_search.py # Semantic Scholar integration for lit-aware optimization
│ │ ├── web_search.py # Web search integration
│ │ ├── workspace_executor.py # Remote workspace command execution
│ │ ├── sandbox_client.py # Remote sandbox HTTP client
│ │ ├── providers.py # Sandbox provider registry (lazy imports)
│ │ ├── backends/ # Execution backends
│ │ │ ├── protocol.py # Backend/SandboxProvider/SandboxClient protocols
│ │ │ ├── worktree.py # Local git-worktree backend
│ │ │ ├── pool.py # Shared local slot pool backend
│ │ │ ├── pool_state.py
│ │ │ ├── remote.py # Remote container sandbox backend
│ │ │ ├── remote_state.py
│ │ │ └── state_keys.py
│ │ ├── hosts/ # Per-host subagent dispatch (claude_fork.py)
│ │ ├── host_install/ # Plugin installers per host
│ │ │ ├── codex.py, claude_code.py
│ │ │ └── _hook_drain.py
│ │ ├── inject/ # Hook injection system (drain, marker, paths, queue, registry)
│ │ ├── static/ # Dashboard web assets (Flask-served SPA)
│ ├── skills/ # Agent skill definitions (SKILL.md per skill)
│ │ ├── research/ # /sva:research — init workspace, build benchmark, run baseline
│ │ ├── optimize/ # /sva:optimize — the main optimization loop
│ │ ├── subagent/ # Internal protocol for optimization subagents
│ │ ├── infra-setup/ # Remote backend provider setup
│ │ ├── literature_search/ # Literature-aware hypothesis generation
│ │ ├── web_search/
│ │ └── references/ # CLI quick-ref, agent SDK reference
│ ├── hooks/hooks.json # Claude Code hooks: PreToolUse, UserPromptSubmit, SessionStart
│ ├── bin/ # Shell wrappers + Rust sva-hook-drain binary
│ ├── .claude-plugin/plugin.json
│ └── .codex-plugin/plugin.json
├── sdk/python/ # Python agent SDK (svatah-agent) — `from sva_agent import Run`
├── sdk/node/ # Node.js agent SDK — `import { Run } from 'svatah-agent'`
├── tests/
│ ├── unit/ # Unit tests (pytest, runs in CI on ubuntu + windows, py3.10-3.14)
│ ├── e2e/ # End-to-end tests (worktree lifecycle, lineage, pool, localhost)
│ └── live/ # Live provider tests (modal, aws, azure, e2b, daytona)
└── scripts/ # Utility scripts (dashboard, scratchpad, version check, graph)
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.
- 11d ago First seen · 167 lines · 2,351 tokens per session scan A d45d1a73752e
Svatah CLAUDE.md is an instructions file published in the GitHub repository bayeslabs-rsi/Svatah (3 stars, last pushed 29d ago), licensed Apache-2.0. It adds 2,351 tokens to every session, about $0.0118 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
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).