Behavior Intelligence · Scoped answers

How is context scoped or inherited across products and agents?

The evidence for all six products preserves meaningful scope boundaries, but the mechanisms differ. Codex receives harness and repository context; the selected Claude evidence is tied to dated chat-product prompts; Claude Code, Gemini CLI, ChatGPT, and Cursor document layered project or user context. These conclusions should not be generalized to every model, plan, surface, or runtime configuration.

Six product answers

What the selected evidence establishes

Keep instructions attached to the correct product and scope instead of assuming one context layer applies everywhere.

Open-source agentic coding interface

OpenAI Codex

Documented

Direct factual answer

Codex's selected sources establish harness-supplied workspace context and authored guidance from Codex home followed by project instruction files. Current discovery documentation is separate from the pinned GPT-5.2 prompt excerpt.

Surface: Codex CLI
Model context: Not established
Configuration / version scope: The current instruction-discovery claim covers Codex CLI; the historical harness excerpt remains scoped to GPT-5.2.
Reviewed: 2026-09-14

Limits and uncertainty

The complete runtime context assembly is not established.

Evidence

Web, desktop, and mobile AI assistant

Claude

Documented

Direct factual answer

The selected Claude evidence applies to dated claude.ai and mobile product prompts and explicitly separates that scope from Claude API behavior.

Surface: claude.ai and Claude apps
Model context: Claude Opus 4.5
Configuration / version scope: Coverage is tied to selected Claude Opus 4.5 entries dated 2025-11-24 and 2026-01-18.
Reviewed: 2026-08-11

Limits and uncertainty

The latest Claude product instruction state is not established.

Evidence

Multi-surface agentic coding product

Claude Code

Documented

Direct factual answer

Claude Code documents managed, user, project, and local CLAUDE.md scopes plus separate repository-scoped auto memory.

Surface: Claude Code terminal CLI · Claude Code IDE integrations
Model context: Not established
Configuration / version scope: The reviewed evidence covers terminal and IDE-oriented product surfaces with a selectable Claude model.
Reviewed: 2026-08-11

Limits and uncertainty

The complete default system prompt is not established.

Evidence

Open-source agentic command-line interface

Gemini CLI

Documented

Direct factual answer

Gemini CLI documents global, workspace, and just-in-time GEMINI.md context plus dynamic injection of enabled skills, subagents, and tools.

Surface: Gemini CLI
Model context: Not established
Configuration / version scope: The product uses a configurable Gemini model and can fully replace its built-in system prompt.
Reviewed: 2026-08-11

Limits and uncertainty

No single immutable runtime assembly is claimed.

Evidence

Multi-surface conversational AI product

ChatGPT

Documented

Direct factual answer

ChatGPT documents custom instructions across conversations, project-bounded instructions and files, separately controlled memory, and selectable model or mode context.

Surface: ChatGPT web · ChatGPT apps
Model context: Not established
Configuration / version scope: Plan, model, surface, enabled tools, and memory settings can change the represented context.
Reviewed: 2026-08-11

Limits and uncertainty

The complete current base prompt and universal runtime configuration are not established.

Evidence

AI code editor, coding-agent platform, and early-beta git forge

Cursor

Documented

Direct factual answer

Cursor documents user rules across projects, version-controlled project and path-aware rules, repository memories, and separately selectable provider models.

Surface: Cursor editor
Model context: Selectable provider models
Configuration / version scope: The reviewed evidence covers editor and CLI surfaces whose permissions can differ.
Reviewed: 2026-08-11

Limits and uncertainty

The complete base system prompt is not established.

Evidence