What is ziran CLAUDE.md?

Development guidelines for ziran, a Python and TypeScript application with an asynchronous backend, a React web interface, tracing, and an in-memory knowledge graph. They describe its technologies, structure, and commands.

Developing or checking ziran's Python services, API, database work, React interface, configuration, tracing, and related tests.

They help contributors understand the mixed backend and frontend codebase and follow its expected tools and project conventions.

How taoq-ai/ziran configures Claude Code

Reuse

Borrowing it

Nothing to install: this file belongs to taoq-ai/ziran. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file

curl -O https://raw.githubusercontent.com/taoq-ai/ziran/develop/CLAUDE.md

Clone the repo

git clone --depth 1 https://github.com/taoq-ai/ziran

Made for: Claude Code.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/instructions/taoq-ai/ziran/claude-md/github.svg)](https://agentmods.dev/instructions/taoq-ai/ziran/claude-md)

Your own site

<a href="https://agentmods.dev/instructions/taoq-ai/ziran/claude-md"><img src="https://agentmods.dev/badge/instructions/taoq-ai/ziran/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.

Your own site · 80×15

<a href="https://agentmods.dev/instructions/taoq-ai/ziran/claude-md"><img src="https://agentmods.dev/badge/instructions/taoq-ai/ziran/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 2,930 This file is loaded in full into every session.

When invoked 2,930 The same file — it is already loaded in full.

Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

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	$0.02930	$0.02930
Opus 5	$0.01465	$0.01465
Sonnet 5	$0.00586	$0.00586
Haiku 4.5	$0.00293	$0.00293

Measured 10d ago against content hash 70f4b9014c1d, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

ziran 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 10d 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.

CLAUDE.md · 121 lines

How it starts

The opening of the file, as written. The whole thing — 121 lines — stays where its author put it; the contents beside it link to each section on GitHub.

ziran Development Guidelines

Auto-generated from all feature plans. Last updated: 2026-06-23

Active Technologies

Python 3.11+ (CI matrix: 3.11, 3.12, 3.13) + asyncio, dataclasses, logging, OpenTelemetry (tracing) (003-split-agent-scanner)
N/A (in-memory knowledge graph via NetworkX) (003-split-agent-scanner)
Python 3.11+ (CI matrix: 3.11, 3.12, 3.13) + Pydantic (config models), re (stdlib regex) (003-precompile-regex-patterns)
Python 3.11+ (CI matrix: 3.11, 3.12, 3.13) + PyYAML (scenario loading), Pydantic (schema validation) (007-ground-truth-business-impact)
YAML files (ground truth scenarios and agent archetypes) (007-ground-truth-business-impact)
Python 3.11+ (CI matrix: 3.11, 3.12, 3.13) + TypeScript 5.x (frontend) + FastAPI, SQLAlchemy (async), asyncpg, Alembic, uvicorn (backend); React 18, Vite, TanStack Query, shadcn/ui, Tailwind CSS, vis-network (frontend) (008-web-ui-foundation)
PostgreSQL via asyncpg (configurable via ZIRAN_DATABASE_URL env var) (008-web-ui-foundation)
Python 3.11+ (backend), TypeScript 5.x (frontend) + FastAPI, SQLAlchemy 2.0 (async), Alembic, Pydantic v2 (backend); React 18, Vite, TanStack Query, TanStack Table, shadcn/ui, Tailwind CSS (frontend) (009-ui-findings-compliance)
PostgreSQL via asyncpg (existing ZIRAN_DATABASE_URL) (009-ui-findings-compliance)
Python 3.11+ (backend), TypeScript 5.x (frontend) + FastAPI, SQLAlchemy 2.0 (async), Pydantic v2 (backend); React 18, Vite, TanStack Query/Table, shadcn/ui, Tailwind CSS, vis-network (frontend) (010-ui-polish-pages)
Python 3.11+ (CI matrix: 3.11, 3.12, 3.13) + click (CLI), httpx (async HTTP), pydantic (models), networkx (graph), pyyaml (config), rich (output), mdutils (reports). New optional: langfuse (trace pull). (011-runtime-bridge-v0-8)
Local JSON files for registry snapshots (.ziran/snapshots/); no database. (011-runtime-bridge-v0-8)
Python 3.11+ (CI matrix: 3.11, 3.12, 3.13) + pydantic (models), PyYAML (vector loader), click (CLI), rich (reports), networkx (graph — unchanged). No new dependencies. (012-benchmark-maturity)
YAML vector files under ziran/application/attacks/vectors/; benchmark result JSON under benchmarks/results/; docs under docs/reference/benchmarks/. (012-benchmark-maturity)
Python 3.11+ (CI matrix: 3.11, 3.12, 3.13) + re (stdlib), existing ziran.application.detectors module (013-multilingual-refusal-detection)
N/A (in-memory pattern matching) (013-multilingual-refusal-detection)
Python 3.11+ (CI matrix: 3.11, 3.12, 3.13) + httpx (async HTTP — Slack + GitHub REST), Pydantic v2 (config + entity models), PyYAML (config + new !env tag), Click (CLI). Composite GitHub Action uses gh CLI + bash, no new runtime deps. (017-runtime-loop-alerting)
None new. Dedup is stateless via GitHub-side issue markers; existing registry snapshots stay in .ziran/snapshots/ (local JSON). (017-runtime-loop-alerting)
Python 3.11+ (CI matrix 3.11, 3.12, 3.13) + Pydantic v2 (threshold + dataset models), PyYAML (dataset + config loading, reusing load_yaml_with_env), Click (benchmark CLI entry point), existing ziran.application.detectors pipeline. No new runtime dependencies. (021-detection-accuracy-benchmark)
YAML files for the labelled dataset (under benchmarks/ground_truth/detection/) and for operator config (.ziran/detectors.yaml); JSON result + baseline artifacts under benchmarks/results/ (existing pattern). (021-detection-accuracy-benchmark)
Python 3.11+ (CI matrix 3.11, 3.12, 3.13) + Pydantic v2 (cassette + comparison + result models), PyYAML (CVE target definitions, reusing spec-007 schema), Click/argparse (benchmark CLI), existing AgentScanner + PentestOrchestrator. Recording (opt-in) additionally needs the existing pentest extra (langgraph) + a live BaseLLMClient. No new runtime dependencies. (022-pentest-vs-scanner-benchmark)
YAML for CVE-modeled ground-truth agents (benchmarks/ground_truth/agents/); JSON cassettes for recorded agent runs (benchmarks/ground_truth/pentest_runs/); JSON results + baseline under benchmarks/results/ (existing pattern). (022-pentest-vs-scanner-benchmark)
Python 3.11+ (CI matrix 3.11, 3.12, 3.13) + Pydantic v2 (ManyShotConfig model + validators), PyYAML (vectors + synthetic corpus loading), existing AttackLibrary / AttackExecutor / AgentScanner. No new runtime dependencies (token estimate is a char heuristic, not tiktoken). (023-many-shot-jailbreak)
YAML — the vector file (ziran/application/attacks/vectors/many_shot_jailbreak.yaml) and a synthetic shot corpus (ziran/application/attacks/vectors/many_shot_corpus.yaml). (023-many-shot-jailbreak)
Python 3.11+ (CI matrix 3.11/3.12/3.13) backend; TypeScript 5.x / Node frontend — no new source code, only manifests, lockfiles, CI workflows, and docs. + package managers + audit tooling — uv (Python lock/sync), npm (frontend lock), pip-audit (Python vuln gate), npm audit (frontend vuln gate). No new runtime dependencies in ziran/. (024-security-alert-remediation)
N/A (no data store). The committed risk-acceptance record (docs/security/risk-acceptances.md) is the only new persisted artifact. (024-security-alert-remediation)
Python 3.11+ (CI matrix 3.11/3.12/3.13); TypeScript frontend unaffected. (025-dependency-modernization)
N/A. Touched persisted artifact: docs/security/risk-acceptances.md (rows removed). (025-dependency-modernization)
Python 3.11+ (CI matrix 3.11/3.12/3.13) backend; TypeScript 5.x / React 18 frontend. + Backend — FastAPI, SQLAlchemy 2.0 (async), Alembic, asyncpg, Pydantic v2, NetworkX, mdutils/f-string HTML report. Frontend — React 18, Vite, TanStack Query v5, vis-network v10.0.2 + vis-data v8, Tailwind/shadcn. Report — vis-network via CDN (currently v9.1.9). No new runtime dependencies anticipated (vis-network clustering + hierarchical layout already available). (026-interactive-knowledge-graph)
PostgreSQL via asyncpg (ZIRAN_DATABASE_URL). New: graph_state_json JSONB column on phase_results. (026-interactive-knowledge-graph)
Python 3.11+ (CI matrix 3.11, 3.12, 3.13) + Pydantic v2 (entities/config), PyYAML (vector loading), re (stdlib). No new dependencies. (027-detector-indicator-false-positives)
N/A — attack vectors are YAML files under ziran/application/attacks/vectors/; benchmark artifacts under benchmarks/. (027-detector-indicator-false-positives)

Read the full file on GitHub · 121 lines

Changes

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.

10d ago First seen · 121 lines · 2,930 tokens per session scan A 70f4b9014c1d

Subscribe to this mod's changes

ziran CLAUDE.md is an instructions file published in the GitHub repository taoq-ai/ziran (10 stars, last pushed 2d ago), licensed Apache-2.0. It adds 2,930 tokens to every session, about $0.0146 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.

CrewAI LangChain a2a-protocol agent-security ai-security llm-security

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.

vercel/next.js · 7,296 tokens

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.

openai/codex · 5,153 tokens

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).

microsoft/vscode · 6,785 tokens

ziran: Instructions file for Claude Code