AGENTS.md
Named by 13 kept items in this workspace.
12 items name it under frameworks and tools and 1 under classification.technologies.
The items that name it 13
File holding pairing rules so both people's agents follow them: branch naming, push after first commit, run partner-files.sh before a task, stop if a planned file is on the list.
Instruction file that coding agents read and trust by default.
Loads AGENTS.md as project instructions.
Built-in Claude Code mod that loads AGENTS.md as project instructions.
Global agent instructions file (claude.md in Claude Code) where persistent security rules such as a dependency cooldown can be set.
File of instructions for the coding agent, which could be introspected and updated but would still drift from the agent's actual behaviour
Repository instructions file for agents; should explain when documents and tests are relevant and authorize safe routine workflows
Generalized equivalent of Claude.md intended to work across different models; Thariq notes he dislikes it but Anthropic will support it since maintaining separate per-model files is a pain.
agent 使用的指令文件,笔记提议从全量加载改为按条件加载并钉住相关片段
emerging cross-tool convention/config file for coding agents, now recognized by Claude Code as an alternative to CLAUDE.md
Instruction file describing agent workflows/behavior for coding agents
Compact version of the Ponytail prompt for agents that read a rules file
In every workspace you can read 24
AI
Security
All technologies in this workspace.