ankyr-improvement-protocol

Flags and implements improvements to skills, rules, and agent instruction files.

1|Updated May 21, 2026
One-click install
npx skills add https://github.com/GuyErreich/AI_Agents --skill ankyr-improvement-protocol-guyerreich
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ankyr-improvement-protocol
Source: https://github.com/GuyErreich/AI_Agents/tree/main/plugins/ankyr-authoring/skills/ankyr-improvement-protocol
Command: npx skills add https://github.com/GuyErreich/AI_Agents --skill ankyr-improvement-protocol-guyerreich

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Reusable patterns, class-of-bug fixes, and instruction gaps discovered during development are often lost because no one documents them. This Skill enforces a mandatory flagging workflow so every improvement opportunity is captured, triaged, and routed to the correct skill or rule file. ## Core Features & Use Cases - Mandatory post-task flag block: Ends every task with an explicit Improvement flags section, even when empty, so nothing is silently dropped. - Structured flag template: Captures category, target file, scope, portability, effort, and priority for each proposed documentation change. - Portable vs project routing: Maps each lesson to the right layer, from portable plugin skills and rules to project-specific AGENT.md chains. - Use Case: After fixing a recurring overlay dismiss lifecycle bug, the agent flags it as a class-of-defect pattern targeting the ankyr-web UX references, then implements the doc update in a parallel session after user approval. ## Quick Start Ask the agent to flag any reusable pattern or instruction gap discovered during the current task using the improvement protocol.

Frequently Asked Questions about ankyr-improvement-protocol

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

FAQPage Schema
How do I document a reusable pattern discovered during coding?▼

End the task with a mandatory Improvement flags block using the flag template, specifying category, target file, description, scope, effort, and priority. The agent then asks whether to implement the change now, in a parallel session, or skip it.

When should I flag an improvement to skills or rules?▼

Flag when a code pattern would benefit from documentation, an instruction is incomplete or outdated, a new architectural or UI pattern should be standardized, or a fix addresses a class of defect even on first occurrence.

Can the agent update skill files without my approval?▼

No. Flagging is automatic and mandatory, but implementing documentation or skill changes requires explicit user approval, preferably handled in a separate parallel session so the main task continues uninterrupted.

Where should portable vs project-specific guidance go?▼

Generic patterns belong in portable plugin skills and rules, while app-specific tokens, component names, and platform details belong in the repo AGENT.md chain or project skills under .cursor/skills/project.

What improvements should not be flagged?▼

Skip one-off typos, single-instance data bugs, changes that contradict existing style, massive refactors, and speculative features. Only flag patterns appearing multiple times or reusable classes of defects.