active-memory-reminder

Extracts session decisions into claude-decisions.json for durable cross-session memory.

754|129|Updated Jun 30, 2026
One-click install
npx skills add https://github.com/Archive228/loopkit --skill active-memory-reminder
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: active-memory-reminder
Source: https://github.com/Archive228/loopkit/tree/main/skills/active-memory-reminder
Command: npx skills add https://github.com/Archive228/loopkit --skill active-memory-reminder

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Context compaction in long-running coding sessions strips the reasoning behind past decisions, causing agents to re-litigate settled design choices and waste token budget. This Skill preserves the 'why' by maintaining a machine-readable decision log alongside the human-readable progress file.

Core Features & Use Cases

  • Dual-file memory model: Pairs free-form claude-progress.txt (narrative context) with claude-decisions.json (structured {ts, decision} entries) so both intent and hard choices survive compaction.
  • Automatic decision capture: A pre-compact hook scans transcripts for decision phrases (decided, chose, rejected, switched from X to Y), dedupes, timestamps, and appends up to 20 entries per compaction without ever blocking the compaction itself.
  • Append-only contract: Entries are never edited or deleted; contradictions are resolved by newer entries with the reason noted in the progress file.
  • Use Case: At the start of a new session after /clear, read claude-decisions.json before touching code to confirm a prior session already rejected a polling approach, avoiding re-deriving that decision.

Quick Start

Read claude-decisions.json alongside claude-progress.txt at session start and check it before proposing any design change.

Frequently Asked Questions about active-memory-reminder

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

FAQPage Schema
How do I preserve decisions across Claude Code compaction?

Maintain a claude-decisions.json file as an append-only array of {ts, decision} entries. A pre-compact hook scans the transcript for phrases like 'decided', 'chose', and 'rejected', then timestamps and appends matching decisions before compaction runs.

What is the difference between claude-progress.txt and claude-decisions.json?

claude-progress.txt holds free-form prose narrative for human-readable context and intent. claude-decisions.json holds structured, timestamped decision records that survive compaction, since models rewrite prose freely but edit structured data carefully.

When should I read claude-decisions.json in a session?

Read it at session start, after /clear, and after compaction, right after claude-progress.txt and before touching code. Also grep it before proposing a design change to avoid re-litigating choices a prior session already settled.

Can I edit or delete old entries in claude-decisions.json?

No, the file is append-only in normal operation. If entries contradict, the newer one wins, and you should note the contradiction in claude-progress.txt so the reasoning is captured rather than erasing the durable record.

What happens if the pre-compact hook misses a decision?

The hook's regex catches most decisions but not all. If you make a decision the transcript will not obviously match, add it manually using the same {ts, decision} shape before ending the session.