Can I run tasks sequentially instead of in parallel?

Yes, by using the --serial flag, you can force the executor to run one item at a time, which is helpful for debugging or resource-constrained environments.

What is the role of FABER Executor compared to FABER Planner?

The planner creates the roadmap and identifies specific work items, while the executor orchestrates the sub-agents (managers) to perform the work and aggregate the final results.

Does FABER Executor clean up temporary files?

Yes, it automatically triggers worktree cleanup for successful items once PRs are merged, ensuring your project directory remains organized.

How does the resume feature handle failed tasks?

The executor stores the state of every item in a JSON file. If a run is interrupted, the --resume flag reads this state and allows the agent to skip completed steps and start exactly where it left off.

FABER Executor

Name: FABER Executor
Author: fractary

$fractary$ byfractary

•

Productivity & Workflow

Executes complex multi-step FABER plans by orchestrating autonomous sub-agents for parallel or serial task completion.

The FABER Executor is a high-reliability orchestration skill designed to transform structured development plans into completed code. It acts as the execution engine for the FABER ecosystem, reading plan files and spawning specialized sub-agents to handle individual work items. Designed for stability, it features parallel execution modes, persistent state logging, and exact-step resume capabilities, ensuring that large-scale codebase migrations or multi-issue tasks are processed efficiently while maintaining a clear audit trail and automatic environment cleanup.

Key Features

01Exact-step resume support for continuing interrupted long-running tasks

02Fail-safe execution logic that continues processing remaining items if one fails

032 GitHub stars

04Parallel and serial execution modes for flexible agent orchestration

05Automatic aggregation of results from multiple background sub-agents

06Integrated worktree cleanup for completed and merged pull requests

Use Cases

01Automating the implementation of a batch of GitHub issues from a single plan

02Managing long-running agentic workflows that require state persistence and recovery

03Executing large-scale codebase refactors across multiple directories simultaneously

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add fractary/claude-plugins faber-executor

For use in Claude.ai and ChatGPT

Download Skill