Skip to content
Back to skills

Goal

ASecurity

Long-running goal tracking. Use when the user gives a complex task requiring multiple steps to complete. Triggers: "goal", "keep going until done", "don't stop until X".

  • 226 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added October 11, 2026
ai-agentsgo

Works with

  • terminal

Security analysis

A100/100

Scanned October 11, 2026

npx -y skills add KonghaYao/peri --skill goal --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Goal?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Goal
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/konghayao-goal/badge)](https://www.skillsdirectory.com/skills/konghayao-goal)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: goal
description: >
  Long-running goal tracking. Use when the user gives a complex task requiring
  multiple steps to complete.
  Triggers: "goal", "keep going until done", "don't stop until X".
userInvocable: true
argumentHint: "[objective description]"
---

# Goal Mode

## When to Use

- The user gives a complex task that requires multiple rounds to complete
- The user says "keep going until done", "don't stop midway", "until X"

## How to Use

1. Call the `goal` tool with action=create and a specific, verifiable objective
2. Work continuously until the goal is achieved
3. Achieved → `goal` tool, action=complete
4. Blocked by something unsolvable → `goal` tool, action=block, reason=why you're blocked
5. Need to abandon/reset the goal → `goal` tool, action=clear (releases the singleton slot, works on terminal states)
6. Check current goal status anytime with action=get

## Important Constraints

- **Objectives must be specific and verifiable**: "improve code" is bad, "raise test coverage to 80%" is good
- **complete is verified**: The system uses an auxiliary LLM to check if you actually achieved the goal. If verification fails, you'll get the reason back
- **block is a distress signal**: Only use when truly unable to continue (missing permissions, missing dependencies)
- **Self-driven after creation**: Once a goal is created, you'll receive a reminder after each turn. You must decide: continue / complete / block / clear
- **Singleton**: Only one goal at a time. Use action=clear to release the slot before creating a new one

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.