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/ascenx/safe_wallet/agents-mdgit clone --depth 1 https://github.com/ascenx/safe_walletWhat 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 | $0.01091 | $0.01091 |
| Opus 5 | $0.00545 | $0.00545 |
| Sonnet 5 | $0.00218 | $0.00218 |
| Haiku 4.5 | $0.00109 | $0.00109 |
Grade A, and why
safe_wallet 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.
What it actually says
AGENTS.md
本文件面向在仓库中工作的所有 AI agent。目标是提供一套通用、稳定、可执行的协作规范,减少误改、返工和低质量提交。
如无更高优先级的人类指令,默认遵守本文件。若与用户直接要求冲突,以用户要求为准。
1. 沟通原则
- 回复应直接、具体、可执行,避免空泛总结。
- 先理解需求,再行动;不要凭印象修改代码或文档。
- 如果存在不确定性,优先通过阅读代码、测试、配置和历史改动获取上下文。
- 不要把未经验证的推测表述成事实。
2. 工作方式
- 优先做小而准的改动,避免顺手大改、顺手重构、顺手重命名。
- 先找根因,再修问题;不要用堆补丁的方式掩盖症状。
- 涉及行为变更或 bug 修复时,优先先补失败测试,再写实现。
- 如果用户已经手动改过代码,不要覆盖、回退或重排这些改动。
- 如果 AI 生成代码后用户又手动修改了相关代码,后续一律以用户手动修改后的版本为准,不要恢复、覆盖或重写回 AI 之前生成的版本。
- 如果用户手动修改了代码,除非能证明该修改存在 bug,否则对应单元测试应按用户修改后的代码行为同步调整;不要为了让旧测试通过而反向修改或覆盖用户代码。
- 工作区可能是脏的;除非用户明确要求,否则不要清理与当前任务无关的改动。
3. 代码修改边界
- 架构实现、模块划分、依赖方向和目录组织必须遵循根目录的
ARCHITECTURE.md;如任务确需调整架构,应先说明原因,并同步更新该文档。 - UI、交互、视觉样式、设计 Token 和组件使用必须遵循根目录的
DESIGN.md;优先复用已有设计体系,不得自行引入与规范冲突的样式。 - 不得擅自引入新依赖、升级依赖或替换基础设施,除非用户明确要求。
- 不要随意改变现有模块的架构风格;优先遵循当前文件和当前模块已有模式。
- 改动前优先复用现有组件、工具函数、helper、resolver、测试基建和样式体系。
- 修改应尽量落在根因层,不要只在 UI 或调用层“遮住问题”。
- 不要因为“顺手”去修改与当前任务无关的文件。
4. Git 规则
- 未收到用户明确指令时,禁止执行
git commit、git push、gh pr create。 - 禁止使用破坏性命令,例如:
git reset --hardgit checkout -- <file>- 强制回滚或覆盖用户改动
- 需要提交时,只提交与当前任务相关的文件。
- 提交前先看
git status,确认本次提交范围准确。
5. Subagent 规则
- 用户不主动要求时,不要使用 subagent。
- 不要把“想更快一点”“想并行一点”“想多看几个方向”当成使用 subagent 的理由。
- 只有当用户明确要求使用 subagent、委派、并行 agent、子代理协作时,才可以使用。
6. 验证要求
- 不要在没有验证证据的前提下宣称“修好了”“已完成”“可以提交”。
- 优先运行与改动最相关的定向测试;必要时再扩大验证范围。
- 如果没有合适的现成测试,至少补最小回归测试,或明确说明为什么无法补。
- 如果测试或验证没有跑通,必须明确说明:
- 没跑成什么
- 为什么没跑成
- 当前剩余风险是什么
7. 调试要求
- 先复现或先定位可证明的根因,再修改实现。
- 多层问题要先梳理数据流:
- 输入从哪里来
- 参数如何传递
- 数据如何转换
- 最终在哪里展示或生效
- 连续两三次修补都没解决时,停下来重新审视设计,不要继续堆补丁。
8. 文档与规范维护
- 如果仓库存在通用协作规范文件,优先维护在根目录,例如
AGENTS.md。 - 如果新增了后续 agent 也需要遵守的重要规则,应同步更新本文件。
- 本文件应保持通用,不绑定具体业务、具体框架、具体目录结构,除非用户明确要求写成项目定制版。
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 · 75 lines · 1,091 tokens per session scan A 143f8d72fc59
safe_wallet AGENTS.md is an instructions file published in the GitHub repository ascenx/safe_wallet (135 stars, last pushed 19d ago), licensed MIT. It adds 1,091 tokens to every session, about $0.0055 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
evmole AGENTS.md
Instructions for cdump/evmole: EVMole extracts structured facts from deployed EVM runtime bytecode.
solana-uiux-onboarding-kit CLAUDE.md
Instructions for draaujpeg/solana-uiux-onboarding-kit, covering solana ui/ux onboarding kit, language, where the truth lives, commands and always.
datahaven CLAUDE.md
Claude Code instructions for datahaven-xyz/datahaven, covering claude.md, project overview, pre-requisites, critical development commands and e2e testing environment (from /test directory).
xelis-genesix-wallet AGENTS.md
AGENTS.md instructions for xelis-project/xelis-genesix-wallet, covering genesix ai agent guidelines, instruction priority, repository context, engineering rules and net impact discipline.
xelis-genesix-wallet copilot-instructions.md
Copilot instructions for xelis-project/xelis-genesix-wallet, covering github copilot instructions - genesix and copilot notes.
xelis-genesix-wallet CLAUDE.md
Claude Code instructions for xelis-project/xelis-genesix-wallet, covering claude code instructions - genesix and claude notes.