# Technical feature reference
[User guide](../user/README.md) · [Documentation home](../README.md) · [Repository README](../../README.md)
This reference defines exact settings, limits, storage behavior, integration contracts, and implementation details. For task-oriented setup and everyday use, start with the [user guide](../user/README.md).
## User feature guides
- [Providers and models](../user/providers-and-models.md)
- [Configuration](../user/configuration.md)
- [Sessions and context](../user/sessions.md)
- [Review and history](../user/review-and-history.md)
- [Customization](../user/customization.md)
- [Side agents and parallel work](../user/agents-and-delegation.md)
- [Tools and safety](../user/tools-and-safety.md)
- [Integrations and automation](../user/integrations.md)
- [Appearance and terminal features](../user/appearance.md)
- [Troubleshooting](../user/troubleshooting.md)
## Getting started
- [Installation and updates](installation-and-updates.md): install with Cargo, build from source, or update.
- [Provider authentication](provider-authentication.md): connect Codex, Anthropic, or a custom provider.
- [Configuration](configuration.md): choose settings, models, and where local data lives.
- [Troubleshooting](troubleshooting.md): fix startup, login, session, and tool problems.
- [Security notes](security.md): protect credentials and understand tool boundaries.
## Everyday use
- [Commands and interfaces](commands-and-interfaces.md): CLI flags, slash commands, and prompt tags.
- [Mission Control](mission-control-tui.md): terminal interface, shortcuts, activity, and queued prompts.
- [Prompt-history suggestions](prompt-history.md): repository-shared inline suggestions, acceptance, and privacy controls.
- [Diff review](diff-review.md): inspect worktree changes, save line comments, and send review context.
- [Side agent](side-agent.md): a separate conversation alongside primary work, with its own history and cancellation.
- [Runtime theming](runtime-theming.md): choose colors and customize appearance.
- [Sessions, context, and cache](sessions-context-cache.md): resume, compact, export, and clean up sessions.
- [Session summarizer](session-summarizer.md): optional progress summaries, per-session controls, and provider settings.
- [Instructions, prompts, skills, and agents](instructions-prompts-skills-and-agents.md): customize how the agent works.
- [Parallel subagents](parallel-subagents.md): delegate work and configure agent profiles.
## Tools and integrations
- [Tools and safety model](tools-and-safety.md): built-in tools, examples, and limits.
- [Code Mode](code-mode.md): JavaScript tool batches enabled by default, compact inventory, limits, and recovery.
- [MCP stdio and HTTP tools](mcp-stdio-tools.md): connect external tool servers.
- [Local-agent stdio protocol](local-agent-protocol.md): shared runtime readiness, owned sessions, bounded history, streamed full-agent turns, approvals and cancellation.
- [Local-agent backend handoff](local-agent-handoff.md): build/setup, captured synthetic traces, tested revision and native/client/platform verification limits.
- [LSP diagnostics](lsp-diagnostics.md): get language-server feedback after edits.
- [Tool-call hooks](tool-call-hooks.md): run local scripts around tool calls and messages.
- [Jev and internal tooling](jev-and-internal-tooling.md): configure skill suggestions, command and prose checks, prompt injection protection, completion verification, and cache handling.
- [Offline tool-output measurement](tool-output-measurement.md): estimate potential token savings from stored sessions without changing live requests.
## Internals and maintainer references
- [Prompt construction](prompt-construction.md): how instructions and context become provider requests.
- [Session replay reuse](session-replay-reuse.md): reuse replay data across requests.
- [Verification](verification.md): local compilation checks and CI policy.
- [Profiling benchmarks](profiling-benchmarks.md): retained workloads, manual runners and measurement limits.
- [Duplicate detection](duplicate-detection.md): find repeated code and documentation.
- [Prompt-cache audit scripts](prompt-cache-audit-scripts.md): static checks and optional live cache checks.
- [CPU hotspot profiling](cpu-hotspot-profiling.md): collect and compare CPU measurements.
- [Memory usage profiling](memory-usage-profiling.md): collect and compare memory measurements.
- [Repeated setup benchmarks](repeated-setup-benchmarks.md): measure startup and repeated setup work.
- [Historical records](../archive/index.md): dated performance measurements, past benchmarks, and release acceptance evidence.
## Editing these docs
Keep technical contracts here: exact behavior, settings, precedence, limits, storage, failure semantics, and source maps. Put task-oriented instructions under `docs/user/`, then link here for advanced detail instead of duplicating contracts.
Keep examples free of real credentials, account ids, auth output, and private session data. Current product behavior and design contracts belong in `docs/user/` and `docs/features/`; historical ADRs, PRDs, and plans live in the [documentation archive](../archive/). See the [documentation guide](../documentation-guide.md) for scope and review rules.