skills.rest
ooooooooooooooooooop
ooooooooooooooooooopCOMMUNITY
Explore all skills by this author →
GitHub

task-mode-router

Classifies tasks by scale and risk to select the appropriate execution depth.

3|Updated Feb 13, 2026
Search Tags:#workflow#planning#code review#agent#task routing#risk classification#execution mode
One-click install
npx skills add https://github.com/ooooooooooooooooooop/personal-ai --skill task-mode-router-ooooooooooooooooooop
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: task-mode-router
Source: https://github.com/ooooooooooooooooooop/personal-ai/tree/main/skills/task-mode-router
Command: npx skills add https://github.com/ooooooooooooooooooop/personal-ai --skill task-mode-router-ooooooooooooooooooop

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? It prevents over-engineering simple requests and under-protecting risky ones by routing every task to the lightest execution mode that still controls its actual risk. ## Core Features & Use Cases - Five-mode classification: Routes requests into small, audit, medium, large, or high-risk modes based on scope, reversibility, and risk. - High-risk gating: Flags installs, global config changes, deletions, permission changes, and browser session usage for audit before any action. - Skill handoff routing: Directs ambiguous requests to clarify-before-change, modifications to minimal-implementation, and large multi-stage work to unified-taskflow. - Use Case: A user asks to refactor a workflow rule across several skill files and update tests; the router classifies it as a large task, requires goals and acceptance criteria first, and hands off to a worker subagent instead of editing directly. ## Quick Start Use task-mode-router to classify my request and tell me the execution mode, allowed actions, and the next step.

Frequently Asked Questions about task-mode-router

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I decide whether a task needs a plan or just a direct edit?▼

Classify the task by scope and risk: single-file, low-risk, easily reversible changes are small tasks done directly, while multi-file or behavior-affecting work needs a short plan with impact and rollback notes.

What counts as a high-risk task in an AI agent workflow?▼

High-risk tasks include installing tools, MCP servers or hooks, modifying global configuration, deleting files, changing permissions, running untrusted code, and using browser login sessions. These require auditing sources, write paths, and rollback before acting.

When should a large task trigger a subagent instead of direct editing?▼

When a task spans multiple files, architecture changes, or needs acceptance criteria, the main session should stop local probing, produce a structured brief, and delegate execution to an independent worker subagent.

Does task classification ever modify files by itself?▼

No. Classification only routes the task; it never writes files. Actual changes are handed to execution skills like minimal-implementation or unified-taskflow after the mode is determined.

Why does a read-only audit request sometimes turn into unwanted changes?▼

Scope creep happens when agents fix issues found during review. The router enforces that read-only requests stay read-only and never expand into syncing, cleaning, or modifying configuration without explicit authorization.

Related Skills

Explore More Software Engineering Skills→
TeXmeijinTeXmeijin

goal-template-generator

Generates execution-ready GOAL templates from rough task requests for AI agents.

Community
Advanced
kaminocorpkaminocorp
1

test-driven-development

Enforces RED-GREEN-REFACTOR test-first workflow for features and bugfixes.

Official
Intermediate
NealschNealsch

Forge-Grill

Clarifies and validates human intent into structured engineering objectives before implementation begins.

Community
Intermediate
coco-researchcoco-research
264

gsd-fast

Execute trivial code changes directly in the current context.

Community
Basic