write-memory

Persists session decisions, conventions, ambiguities, and issues into structured .ai-memory files.

Updated May 16, 2026
One-click install
npx skills add https://github.com/alimtvnetwork/img-pdf-v2 --skill write-memory-alimtvnetwork
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: write-memory
Source: https://github.com/alimtvnetwork/img-pdf-v2/tree/main/.agents/skills/write-memory
Command: npx skills add https://github.com/alimtvnetwork/img-pdf-v2 --skill write-memory-alimtvnetwork

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI coding sessions lose context between turns, forcing the next session to rediscover decisions, conventions, and open questions. This Skill enforces a disciplined write-to-memory workflow so every decision, plan change, ambiguity, and bug is recorded in a structured .ai-memory/ directory before the turn ends. ## Core Features & Use Cases - Memory Routing Protocol: Routes new information to the correct destination — learned conventions, strictly-avoid rules, suggestions, issues, plans, or ambiguity files. - Index Synchronization: Requires every new memory or plan file to be registered in its corresponding index file in the same operation, keeping the knowledge base navigable. - Ambiguity Lifecycle Management: Tracks open questions in a dedicated folder and moves them to a resolved folder with a Resolution block once answered. - Use Case: After a long refactoring session where several conventions were agreed upon and one requirement remained unclear, run this Skill to log the conventions, record the open ambiguity, update the plan index, and produce a completion summary for the next session. ## Quick Start Ask the AI to persist everything decided and discovered this session into .ai-memory using the write-memory rules.

Frequently Asked Questions about write-memory

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

FAQPage Schema
How do I persist AI session context between conversations?

Write decisions, conventions, and open questions into a structured .ai-memory directory with indexed markdown files. Each memory file is registered in an index so the next session can reload context without guessing.

How do I track unresolved ambiguities in an AI coding workflow?

Store each open question as a separate file in an ambiguity folder. When the user answers, move the file to a resolved folder and append a Resolution block, keeping a clean audit trail of decisions.

What is the best way to log coding conventions discovered during a session?

Record each convention as an institutional memory file under a learned/ folder and register it in the memory index in the same operation. Append-only rule files capture things that must never happen again.

Can this Skill modify application source code while writing memory?

No. The Skill explicitly forbids any execution beyond writing to .ai-memory, fixing root readme casing, and moving resolved ambiguity files. Application source changes are out of scope for the memory turn.

Why does the Skill require a git log audit before writing memory?

Running git log on the last 30 commits lets the AI summarize recent progress and extract lessons learned, ensuring memory entries reflect actual repository history rather than assumptions.