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 rules/divar-ir/ai-doc-gen/code-patternsgit clone --depth 1 https://github.com/divar-ir/ai-doc-genWhat 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 | $0.00000 | $0.00636 |
| Opus 5 | $0.00000 | $0.00318 |
| Sonnet 5 | $0.00000 | $0.00127 |
| Haiku 4.5 | $0.00000 | $0.00064 |
Grade A, and why
code-patterns 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 — 55 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Code Patterns
Style
- Ruff: 120-char lines, 4-space indent,
target-version = "py313"; import sorting enabled — runuv run ruff format src/ && uv run ruff check src/. - Type hints everywhere;
pathlib.Pathfor all file paths (never strings); PydanticBaseModelfor all config/data structures. - Naming: snake_case files/functions, PascalCase classes,
_privatemethods, UPPER_SNAKE constants, config classes end inConfig.
Configuration
class MyHandlerConfig(BaseHandlerConfig, MyAgentConfig):
exclude_feature: bool = Field(default=False, description="Exclude feature") # description feeds CLI --help
- Every
Fieldneeds adescription— CLI arguments are generated from it (--exclude-feature, store_true, defaultNone= "not specified"). - Load order: Pydantic defaults <
.ai/config.yaml(dot-notation section keys) < CLI args, merged withmerge_dicts(). - Env vars in
src/config.py: required →os.environ["KEY"]; optional →os.getenv("KEY", "default")withint()/float()/str_to_bool()conversion.
Async and concurrency
- All handler/agent operations are async; handlers implement
async def handle(self). - Bounded concurrency: use
WorkerPoolfromsrc/utils/worker_pool.py(WorkerPool(max_workers=0)= CPU count) for many tasks;asyncio.gather(*tasks, return_exceptions=True)for a fixed small set. - Always isolate errors: check each result with
isinstance(result, Exception), log withexc_info=True, continue.
Error handling
- Partial success is acceptable: warn and continue if some agents fail; raise
ValueErroronly on complete failure. - Guarantee cleanup with try/finally (e.g., cronjob project cleanup).
- Inside agent tools, raise
ModelRetry(from pydantic_ai) for recoverable errors — file not found, permission denied.
Logging and tracing
from utils import Logger
Logger.init(logs_dir) # once per execution, before any use
Logger.info("Agent completed", {"total_tokens": usage.total_tokens}) # structured dict payload
Logger.error("Failed", exc_info=True) # always exc_info=True for errors
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 · 55 lines · 0 tokens per session scan A dbf6d957d057
code-patterns is a cursor rule published in the GitHub repository divar-ir/ai-doc-gen (752 stars, last pushed 1mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 636 tokens. 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 cursor rules, from other repositories
desktop-commander-default
Prefer Desktop Commander MCP for local filesystem, terminal, process, search, and workspace operations.
conversation-logs
仅当用户贴 32-hex / UUID / 排查包,或明说查产品 AI 失败 / 巡检时读。讨论产品或架构默认不打开。用 logs/dev.jsonl + DB,禁止在文件系统搜文件名。.
cross-platform-frontend
前端跨端架构——产品页单树 + 壳自适应;能力黑名单减法;fold 只留桌面一套;必过 conformance.
doc-governance
文档治理——docs/ 与 rules/ 定位、内容策略、同步规则、命名篇幅。写/改 docs/ 或 .cursor/rules/ 前阅读。.
frontend-preview
改桌面前端组件/样式/布局后想自检渲染效果、或想离线查看某个 AI 态(辩论/派单/升级/审批/计划复核/工具/引用…)时读——#/preview 离线回放 conformance 向量、帧滑块逐帧看流式中间态、pnpm dev:web 纯浏览器跑渲染层、pnpm shoot 无头截图自检、协作图两入口视口探针 shoot:graph-probe、CI 渲染冒烟门禁。别再靠跑真实 AI 看前端效果。.
verify-scope
改完代码要跑验证时读——三档验证阶梯的窄化命令配方(pytest ::/-k/--lf、vitest 点名文件/-t/--related、shoot 场景过滤)、release:gate 五段分拆与失败续跑、反空转与跨包边界。跑 release:gate / 全量 pytest / 全量 vitest / pnpm shoot 前必读。.