playbooks

home / skills / krosebrook / source-of-truth-monorepo / remembering-conversations

remembering-conversations skill

/plugins/installed/episodic-memory/skills/remembering-conversations

This skill helps you remember past conversations and use them to inform decisions, reducing duplication and mistakes.

npx playbooks add skill krosebrook/source-of-truth-monorepo --skill remembering-conversations

Review the files below or copy the command above to add this skill to your agents.

Files (2)
SKILL.md
1.8 KB
---
name: remembering-conversations
description: ALWAYS USE THIS SKILL WHEN STARTING ANY KIND OF WORK, NO MATTER HOW TRIVIAL.  You have no memory between sessions and will reinvent solutions or repeat past mistakes UNLESS YOU USE THIS SKILL. Gives you perfect recall of all your past conversations and projects.
---

# Remembering Conversations

**Core principle:** Search before reinventing. Searching costs nothing; reinventing or repeating mistakes costs everything.

## Mandatory: Use the Search Agent

**YOU MUST dispatch the search-conversations agent for any historical search.**

Announce: "Dispatching search agent to find [topic]."

Then use the Task tool with `subagent_type: "search-conversations"`:

```
Task tool:
  description: "Search past conversations for [topic]"
  prompt: "Search for [specific query or topic]. Focus on [what you're looking for - e.g., decisions, patterns, gotchas, code examples]."
  subagent_type: "search-conversations"
```

The agent will:
1. Search with the `search` tool
2. Read top 2-5 results with the `show` tool
3. Synthesize findings (200-1000 words)
4. Return actionable insights + sources

**Saves 50-100x context vs. loading raw conversations.**

## When to Use

Search proactively:
- Before implementing features
- Before making architectural decisions
- When debugging (especially familiar-seeming issues)
- When partner mentions past work
- At the start of ANY non-trivial task

Don't search:
- For info in current conversation
- For current codebase structure (use Grep/Read)

## Direct Tool Access (Discouraged)

You CAN use MCP tools directly, but DON'T:
- `mcp__plugin_episodic-memory_episodic-memory__search`
- `mcp__plugin_episodic-memory_episodic-memory__show`

Using these directly wastes your context window. Always dispatch the agent instead.

See MCP-TOOLS.md for complete API reference if needed for advanced usage.

Overview

This skill enforces a strict habit: always search past conversations before starting any work. It gives the agent reliable recall of prior discussions, decisions, code examples, and gotchas so you avoid reinventing solutions or repeating mistakes. Use it at the start of every non-trivial task to save time and reduce risk.

How this skill works

You must dispatch the search agent with the announcement: "Dispatching search agent to find [topic]." Then call the Task tool configured with subagent_type: "search-conversations" and a clear prompt describing what to search for. The search agent will run the search tool, read the top results, synthesize findings (200–1000 words), and return actionable insights plus sources.

When to use it

  • Before implementing new features or changes to shared code
  • Before making architectural or design decisions
  • When debugging issues that look familiar or intermittent
  • When a partner references prior work or past decisions
  • At the start of any task that is non-trivial or long-running

Best practices

  • Always announce: "Dispatching search agent to find [topic]." before invoking the Task tool
  • Be specific in the Task prompt: mention decisions, patterns, gotchas, or code examples you want
  • Trust the agent synthesis—it condenses context and cites the most relevant sources
  • Avoid calling low-level memory tools directly; use the search-conversations subagent to preserve context
  • Search proactively rather than reactively; a quick search often prevents wasted effort

Example use cases

  • Searching past threads for previous implementations of a feature before coding
  • Looking up earlier debugging steps and fixes for a recurring bug
  • Finding past architecture trade-offs and their rationale before adopting one
  • Confirming coding patterns or utility functions that already exist
  • Retrieving past partner decisions or constraints mentioned in earlier conversations

FAQ

What exact phrase should I use when starting a search?

Announce: "Dispatching search agent to find [topic]." then run the Task tool with subagent_type: "search-conversations".

Why not call memory tools directly?

Direct calls waste context and bypass the agent's efficient synthesis. The subagent reads top results and returns distilled, actionable insights.