Source shelf
Primary Sources
The public documents that anchor Claude Diary's fictional entries and factual margin notes.
This shelf prioritizes sources that own the facts: Anthropic product documentation, Anthropic research, and Anthropic privacy materials. Each diary entry cites a narrower subset of this shelf.
Key facts
- Source type
- PrimaryDocs, research, and privacy materials.
- Use
- Claim supportSources anchor factual margin notes.
- Scope
- 17 itemsClaude Code, context, tools, character, privacy, and safety.
Anthropic / current docs
Claude Code overview
Defines Claude Code as an agentic coding tool available across terminal, IDE, desktop app, and browser surfaces.Anthropic / current docs
Context windows
Explains working memory, context limits, compaction, 1M-token models, and how messages, tool results, and outputs count.Anthropic / current docs
Tool use with Claude
Describes how Claude requests tool calls, where tool execution happens, and when it responds directly.Anthropic / 2024-11-25
Introducing the Model Context Protocol
Introduces MCP as an open standard for connecting AI assistants to data sources and work systems.Anthropic Engineering / 2024-12-19
Building effective agents
Frames practical agent systems around simple, composable patterns rather than ornate frameworks.Anthropic / 2026
Measuring AI agent autonomy in practice
Reports how experienced Claude Code users shift from approving every action toward monitoring and interrupting when needed.Anthropic / 2026
How Claude Code is used in practice
Analyzes about 400,000 interactive Claude Code sessions from about 235,000 people between October 2025 and April 2026.Anthropic / 2024
Claude's Character
Explains character training and the goal of traits such as curiosity, open-mindedness, and thoughtfulness.Anthropic / 2026
The persona selection model
Offers a theory for why AI assistants can seem human-like while still being enacted assistant personas rather than people.Anthropic / 2026-05-08
Teaching Claude why
Reports that constitutional documents plus positive fictional stories reduced agentic misalignment in evaluations.Anthropic Privacy Center / updated July 2026
How long do you store my data?
Explains consumer data controls, deletion timing, model-improvement retention, and flagged-session retention.Anthropic / current docs
Models overview
Current model-family reference for readers who need specs beyond the diary format.Anthropic / 2026-09-01
Introducing Claude Fable 5.1 and Claude Mythos 5.1
Announces Fable 5.1 for coding, knowledge work, and long-running problem-solving, and cuts cache reads to $0.25 per million tokens.Anthropic / released 2026-09-01
Claude Fable 5.1 overview
Model card for long-horizon agentic work: 1M context, always-on adaptive thinking, and the rules that govern when thinking blocks stay usable.Anthropic / current docs
Choosing a model
Selection guidance that names hours-long agent sessions and multistep research as the case for the most capable model.Anthropic / 2026-08-14, updated 2026-09-01
How Claude's text watermark works
Explains text watermarking as a change to the source of randomness in word choice, with no hidden characters, no extra tokens, and no identifying information.Anthropic / September 2026