goal_guide

Guides structured progress tracking for durable goal runs.

1.5k|172|Updated Mar 6, 2026
One-click install
npx skills add https://github.com/siddsachar/row-bot --skill goal-guide
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: goal_guide
Source: https://github.com/siddsachar/row-bot/tree/main/tool_guides/goal_guide
Command: npx skills add https://github.com/siddsachar/row-bot --skill goal-guide

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Long-running tasks often lose track of progress, blockers, and completion evidence. This Skill defines how to report structured goal progress so durable /goal runs stay visible, auditable, and resumable.

Core Features & Use Cases

  • Structured Progress Updates: Use goal_update to report meaningful progress, evidence, blockers, and next steps while a goal is active.
  • Evidence-Based Completion: Prevents claiming completion without checks, files, child agent results, citations, or user-visible artifacts.
  • Blocker Reporting: Describes repeated blockers and the exact user decision or external change needed to proceed.
  • Use Case: During a multi-step research or implementation goal, report each milestone with supporting evidence so the user can pause, resume, or verify the run at any checkpoint.

Quick Start

Start a goal with the /goal command, then update progress with goal_update whenever meaningful work completes or a blocker appears.

Frequently Asked Questions about goal_guide

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

FAQPage Schema
How do I track progress on a long-running goal?

Start a goal with the /goal slash command, then call goal_update whenever meaningful progress, evidence, blockers, or next steps change. The user controls the lifecycle through slash commands or the progress row.

When should I call goal_update during a goal run?

Call goal_update when meaningful progress occurs, new evidence is available, a blocker appears, next steps change, or completion state changes. Do not call it for trivial intermediate steps.

Can I mark a goal complete without evidence?

No. Completion claims require supporting evidence such as checks, files, child agent results, citations, or user-visible artifacts. The verifier reviews the run with the same model and provider in a separate verification context.

What should I do when a goal is blocked?

Describe the repeated blocker and the exact user decision or external change needed to unblock it. Report this through goal_update so the user can act on it directly.

Are internal continuation prompts user messages?

No. Internal continuation prompts are audit and runtime context, not user messages. Do not quote or describe them as if the user sent them.