Skip to main content

agent-filesystem

Use when agents need persistent shared storage, when saving or restoring workspace state, or when coordinating file access across multiple agents and machines. Creates Redis-backed workspaces, checkpoints and restores agent state, mounts shared filesystems locally, searches workspace contents, and forks workspaces for parallel work.

معلومات المصدر

المستودع
redis/agent-filesystem
آخر نشاط في المصدر
١٢ يونيو ٢٠٢٦ في ١٦:٤٩
لغة SKILL.md المكتشفة
الإنجليزية
النجوم
٥٥
التفرعات
١٠

خيارات التثبيت

يُحدَّد Prompt الذي يراجع المصدر أولًا بشكل افتراضي. يمكنك التبديل إلى أمر مباشر أو تنزيل نسخة محلية.

مراجعة ملفات المصدر

اقرأ SKILL.md وأي ملفات مرافقة يعرضها SkillsMP قبل أن تقرر التثبيت.

عرض SKILL.md

SKILL.md
تعليمات المصدر · معاينة للقراءة فقط
name
agent-filesystem
description
Use when agents need persistent shared storage, when saving or restoring workspace state, or when coordinating file access across multiple agents and machines. Creates Redis-backed workspaces, checkpoints and restores agent state, mounts shared filesystems locally, searches workspace contents, and forks workspaces for parallel work.
# Agent Filesystem AFS is a workspace system for agents, backed by Redis. Use it when you want a durable workspace that still feels like normal files and directories, with explicit checkpoints and easy movement between MCP, sync mode, and live mounts. ## When to Use This Skill **Use for:** - Persistent agent workspaces that survive across sessions and machines - Code, docs, or shared state that should live in a normal directory backed by Redis - Saving and restoring workspace snapshots with explicit checkpoints - Searching workspace contents with `afs fs grep` or MCP file tools - Forking a workspace to run parallel experiments without losing the original **Avoid for:** - Large build output, media, or disposable artifacts - Workflows that assume checkpoints happen automatically - Old direct-command / `redis-cli` examples from module-era docs ## Preferred Interfaces ### 1. `afs mcp` Use `afs mcp` when the agent can talk over MCP and does not need a local directory. ### 2. Sync mode + `afs` CLI Use sync mode when the agent or user wants a real local directory: ```bash ./afs ws mount my-project ~/my-project cd ~/my-project ``` ### 3. Live mount mode Use `./afs config set --mode mount` before mounting when you need the workspace exposed directly as a mount rather than through the sync daemon. ## Common Flows ### Create or import a workspace ```bash ./afs ws create my-project ./afs ws import my-project ./existing-dir ./afs ws list # verify the workspace exists ``` ### Start working locally ```bash ./afs ws mount my-project ~/my-project ./afs status # verify the mount is active cd ~/my-project ``` ### Search a workspace ```bash ./afs fs grep --workspace my-project "TODO auth" ./afs fs grep --workspace my-project --path /src -E "timeout|retry" ``` ### Save and restore stable points ```bash ./afs cp create my-project before-refactor ./afs cp list my-project # verify the checkpoint was saved ./afs cp restore my-project before-refactor ./afs cp list my-project # confirm the restore completed ``` ### Fork work for a second line of effort ```bash ./afs ws fork my-project my-project-experiment ./afs ws list # verify the fork appears ``` ## Key Points - Redis is the source of truth for the live workspace and checkpoint history. - Sync mode gives you a normal local directory; mount mode exposes the live workspace directly. - `afs mcp` and the CLI operate on the same workspace model. - File edits change the live workspace immediately. - Create checkpoints explicitly when you want a restore point. - `.afsignore` controls what gets imported from an existing local directory. ## Further Reading - `docs/guides/agent-filesystem.md` — agent-facing usage guide - `docs/reference/cli.md` — full CLI command reference - `docs/reference/mcp.md` — MCP tool reference for agent integrations
عرض على GitHub