Files
dotfiles/AGENTS.md
T
ddadminandClaude Opus 4.6 3c690bdb12 fix(skills): ужесточены permissions и добавлен TODO для codex-review
- Зачем:
  - Codex-ревью выявило permission bypass через Bash(timeout *) и
    сохраняющиеся safety prompts на $() и > в bash-командах.
- Что:
  - permission сужен до Bash(timeout 600 codex exec *), rm-паттерны конкретизированы.
  - REVIEW_ID использует $RANDOM для энтропии, resume пишет в файл вместо pipe.
  - добавлен TODO в AGENTS.md: план замены bash-паттернов на нативные инструменты.
- Проверка:
  - /codex-review при следующем использовании.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-03-31 10:59:26 +03:00

4.0 KiB

GEMINI.md - Project Context & Instructions

This repository contains personal dotfiles and configurations for a highly productive development environment, integrating Bash, Tmux, Vim, VS Code, and various AI assistants (Gemini CLI, Claude, Codex, Opencode).

📂 Project Overview

  • Type: Non-Code Project (Dotfiles / Configuration)
  • Management: Uses stow for Linux/WSL and chezmoi for Windows/WSL.
  • Key Focus:
    • Unified configuration across editors and terminals.
    • Deep integration with AI tools via shared "skills".
    • Automated environment setup (proxy handling, secret loading).

🛠 Key Components

1. Shell & Terminal (Bash & Tmux)

  • Bash: Custom prompt (.bash_prompt), utility functions (.bash_functions like proxy_on/off), and core customizations (.bash_custom).
  • Tmux: Dual prefix (C-b and C-Space), Vi-mode for copying, and optimized pane management.
  • Secrets: Managed via ~/.env.secrets (not tracked in Git).

2. Editors (Vim, VS Code, Zed)

  • Vim: Minimalist .vimrc, 2-space indentation, leader-key mappings (Space).
  • VS Code (WSL): Uses server-env-setup to automatically load secrets and configure proxy settings upon startup.
  • Zed: Placeholder directory for future configuration.

3. AI Agents & Skills

  • Shared Skills: Located in agents/.agents/skills/. This is the single source of truth for AI instructions.
  • Gemini CLI: Configured via gemini/.gemini/.
    • Policies: readonly-commands.toml allows critical commands like git status, git diff, and git add to run without confirmation.
  • Opencode: MCP-based assistant configuration (opencode.json) with support for documentation, search, and web-reading.
  • Skill Bridges: Symlinks are used to share skills across Claude, Codex, and Opencode.

🚀 Usage & Maintenance

Deployment

  • Linux/WSL: Use stow <package_name> from the root directory.
  • Windows: Use chezmoi apply.

Adding/Updating Skills

  1. Modify files in agents/.agents/skills/.
  2. Ensure symlinks in other tool directories are updated if new skills are added.

Git Workflow for AI (Gemini CLI)

  • Safe Commands: git status, git diff, and git add are pre-approved in the policy.
  • Commits: Always use the conventional-commits skill for generating commit messages. Use the Способ B (temporary file) if direct Heredoc causes issues.

📋 TODO

codex-review: убрать permission prompts

Скилл codex-review (agents/.agents/skills/codex-review/SKILL.md) вызывает safety warnings Claude Code на $() и > в bash-командах. Allow-list не помогает — это отдельный слой проверок.

Решение: переписать шаги скилла, заменив bash-паттерны на встроенные инструменты:

  • Шаг 2 (Session ID): Claude генерирует ID сам, без $(date) в bash.
  • Шаг 3 (Diff): git diff → результат через Write tool вместо > redirect.
  • Шаг 7 (Resume): stdout tool result вместо > file 2>&1.
  • Шаги 4 и 9 — без изменений (-o и rm -f работают).

⚠️ Important Files for Reference

  • README.md: Detailed installation and usage instructions.
  • bash/.bash_functions: Proxy and environment management functions.
  • gemini/.gemini/policies/readonly-commands.toml: Permissions for Gemini CLI.
  • agents/.agents/skills/conventional-commits/SKILL.md: Rules for commit formatting.

🤖 AI Interaction Mandates

  • Adhere to Local Patterns: When suggesting changes, follow the 2-space indent rule and minimalist style established in .vimrc and .bash_custom.
  • Privacy First: Never log or commit contents of ~/.env.secrets.
  • Verification: Before updating configs, verify with git status and git diff (which are allowed without confirmation).