goal-anchor
CommunityAnchor objectives to prevent agent drift.
Software Engineering#multi-turn#agent-prompt#phase-transition#prompt-injection-prevention#drift-prevention#goal-anchor
AuthorMinhThang1009
Version1.0.0
Installs0
System Documentation
What problem does it solve?
Prevents goal drift in long-running or multi-phase AI agent workflows by ensuring every prompt starts with a clearly defined OBJECTIVE.
Core Features & Use Cases
- Injects an OBJECTIVE block at the start of the subagent prompt to anchor the intended goal.
- Enforces drift checks every 5 turns and provides a phase-transition protocol with compaction to preserve objective integrity.
- Supports different thresholds for multi-agent reviews and general tasks, and includes a structured, repeatable process for resuming after interruptions.
Quick Start
Inject the OBJECTIVE block at the start of the subagent prompt before spawning a multi-turn agent.
Dependency Matrix
Required Modules
None requiredComponents
Standard package💻 Claude Code Installation
Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.
Please help me install this Skill: Name: goal-anchor Download link: https://github.com/MinhThang1009/dotclaude/archive/main.zip#goal-anchor Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
Agent Skills Search Helper
Install a tiny helper to your Agent, search and equip skill from 510,000+ vetted skills library on demand.