# code-kb
> **Interactive Showcase & Benchmarks:** [https://anortham.github.io/code-kb/](https://anortham.github.io/code-kb/)
`code-kb` is a fast, lightweight code-intelligence engine and Model Context Protocol (MCP) server designed specifically for AI coding agents.
Backed by the rich AST fact tables produced by [`julie-extractors`](https://github.com/anortham/julie-extractors), `code-kb` provides progressive disclosure, semantic symbol navigation, and surgical context slicing—enabling agents to navigate and understand codebases with **80–90% fewer tokens** without burning context on full file reads or noisy text grep. Agents edit with their native filesystem tools; code-kb automatically refreshes indexed files after filesystem changes.
---
## Why code-kb?
Traditional AI coding agents burn massive amounts of context loading entire source files into their prompt windows just to inspect a single type signature or implementation detail.
`code-kb` solves this with a **progressive disclosure** architecture:
1. **Repository Orientation (`codebase_outline`):** Understand directory structures and key exports in a few hundred tokens.
2. **File Skeletons (`file_skeleton`):** Inspect function signatures, types, traits, and docstrings with implementation bodies stripped.
3. **Symbol Lookup & Discovery (`lookup_symbol` / `search_symbols`):** Instant exact/prefix identifier lookups and conceptual search over names, signatures, and docstrings; substrings inside identifiers are found (`sha256` finds `parseSha256Sidecar`).
4. **Surgical Symbol Context (`get_symbol_context`):** In a single turn, fetch a target function's body along with its callee signatures, parameter types, and associated unit tests.
5. **Fresh Indexes:** Filesystem edits automatically refresh indexed files, and target body and skeleton reads refresh their file before answering.
---
## Key Principles
- **Small Retained Memory:** Written in Rust, zero heavy runtimes (no web dashboard, no GPU models), retained process memory is about 25 MB. The plugin launcher is a small Node script that replaces itself with the native binary via `process.execve` on supported POSIX Node 22+ runtimes (or waits via `spawn` on Windows/older Node).
- **Sub-5ms Query Latency:** Direct SQLite queries in WAL mode with zero in-memory heap bloat.
- **One `project_root` on Every Call:** Every tool except `telemetry_summary` takes `project_root`, the absolute path of the project or git worktree the agent works in (`lookup_symbol(project_root="/path/to/project", query="...")`). A root in every call cannot go stale, so a move to a worktree never answers from the old index.
- **CLI-First Parity:** Every MCP tool has an exact 1:1 CLI command for instantaneous terminal verification and dogfooding.
- **Continuous 3-Tier Sync:** Tool-driven updates, JIT staleness guards before reads, and a debounced background watcher with a Git storm circuit breaker.
---
## Install
Install the plugin for your agent. The plugin ships a small Node.js launcher. On the first
run it downloads the matching `code-kb` release archive for your platform, verifies its
SHA-256, and unpacks `code-kb` and `julie-extract` into `~/.code-kb/dist/<version>/`. Later
runs start instantly. Nothing else to download or put on `PATH`.
Requirements:
- Node.js 18 or newer on `PATH` (the launcher is a Node script; the server itself is a native binary).
- Windows 10 build 17063 or newer (the launcher unpacks with the built-in `tar.exe`).
### Claude Code
```text
/plugin marketplace add anortham/code-kb
```
```text
/plugin install code-kb@code-kb
```
*(Send as two separate prompts.)* The plugin registers the MCP server, the progressive
disclosure skill, a `/code-kb:code-kb-telemetry` command for usage and token-savings reports, a
`/code-kb:report-issue` command that files a bug with a diagnostic bundle, and
SessionStart and SubagentStart hooks that inject routing instructions.
### Codex
```bash
codex plugin marketplace add anortham/code-kb
codex plugin add code-kb@code-kb
```
Run `codex`, open `/hooks`, and trust the two code-kb hooks.
To update, refresh the marketplace snapshot first. `codex plugin remove` followed by
`codex plugin add` reinstalls the version already in the snapshot:
```bash
codex plugin marketplace upgrade
codex plugin remove code-kb@code-kb
codex plugin add code-kb@code-kb
```
### Antigravity CLI (AGY)
```bash
agy plugin install https://github.com/anortham/code-kb
```
The plugin registers the MCP server and a `PreInvocation` hook. Antigravity has no
`SessionStart` hook, so routing directives arrive per turn through `injectSteps`.
### Grok CLI
```bash
grok plugin install anortham/code-kb --trust
```
### First Run & Automatic Indexing
You do not need to run `code-kb scan` by hand. The first `code-kb` tool call or CLI command in
a repository creates `<workspace>/.code-kb/artifact.db` and runs the initial scan. A git
worktree copies its parent repository's index instead of scanning again. Files that changed
while no session ran, for example after a branch switch, are reconciled before the first
answer. Pre-index a large repository before a session with:
```bash
code-kb scan
```
Launcher environment variables:
| Variable | Effect |
|---|---|
| `CODE_KB_HOME` | Directory that holds `dist/` (default `~/.code-kb`). |
| `CODE_KB_VERSION` | Release version to fetch instead of the plugin's own version. |
| `CODE_KB_BIN` | Run this binary and skip the download entirely (local builds). |
Two overrides need no environment variable, which matters in harnesses that do not pass the
environment to MCP servers (Codex):
- A binary or symlink at `~/.code-kb/bin/code-kb` (`code-kb.exe` on Windows) runs instead of any
download, in every harness. For development: `ln -s /path/to/code-kb/target/release/code-kb ~/.code-kb/bin/code-kb`.
- A plugin installed from a source checkout runs that checkout's `target/release/code-kb` when it
exists.
With either in place, `cargo build --release` plus a session restart is the whole development loop.
### Uninstall
| Harness | Command / Action |
|---|---|
| Claude Code | `/plugin remove code-kb` |
| Codex | `codex plugin remove code-kb@code-kb` |
| Antigravity (AGY) | `agy plugin uninstall code-kb` |
| Grok CLI | `grok plugin uninstall code-kb` |
Delete `~/.code-kb/dist` to remove the downloaded binaries. Each workspace keeps its index
in `<workspace>/.code-kb/`; delete that directory to remove the index.
---
## Manual Configuration
Use this path for harnesses without a plugin manager, or when you would rather manage the
binary yourself.
### Step 1: Get the Binaries
- **GitHub Releases:** download the archive for your platform from
[GitHub Releases](https://github.com/anortham/code-kb/releases):
- Linux x86_64 and ARM64 (`.tar.gz`)
- macOS Apple Silicon and Intel (`.tar.gz`)
- Windows x86_64 and ARM64 (`.zip`)
Unpack it and put both binaries on your `PATH` (for example `~/.local/bin`,
`/usr/local/bin`, or `C:\tools`). `code-kb` and `julie-extract` are packaged side by side,
and `code-kb` finds `julie-extract` next to its own executable.
- **Cargo:**
```bash
cargo binstall code-kb-cli
# or
cargo install code-kb-cli
```
Cargo installs `code-kb` only. Download the pinned `julie-extract` from the
[julie-extractors releases](https://github.com/anortham/julie-extractors/releases) (version
in `scripts/julie-pins.json`) and put it on your `PATH` or set `JULIE_EXTRACT_BIN`.
- **Verify:**
```bash
code-kb --version
```
### Step 2: Register the MCP Server
Every harness runs the same command: `code-kb serve`. Hooks run `code-kb hook <Event>`.
#### How code-kb Finds Your Project
Every tool except `telemetry_summary` takes a required `project_root`: the absolute path of the
project or git worktree the agent works in. The agent sends the same value on every call. It
changes the value when it moves to a worktree or another project. The routing hook and the skill
tell the agent to do this.
- A plain path or a `file://` URI is accepted. A relative value is an error.
- A subfolder or a file inside the project resolves to the enclosing project.
- `code-kb` refuses a filesystem root (`/`, `C:\`), the home directory, and a folder with no
project marker (`.git`, `Cargo.toml`, `package.json`, `go.mod`, `pyproject.toml`) and no
code-kb index. The error names the path and the reason, and `code-kb` creates nothing.
- `path` and `file_path` are relative to `project_root`, or absolute inside it. An absolute path
outside `project_root` is an error. A relative or absolute path inside a nested git worktree or
submodule counts as outside `project_root`. The error names that nested root. A path never switches the
project.
- A project with no index gets one on the first call. A git worktree copies its parent
repository's index; any other project runs a full scan. If the index is not ready in 5 s, the
answer is `Indexing <root> started; call again in a few seconds.`
The server keeps one active index. A call for another project switches to that project's
`<root>/.code-kb/artifact.db`.
`--root` on the `serve` command is optional for every client, including GUI apps. It only names
the startup pre-warm root: the server prepares that project's index at start. Without `--root`,
the server pre-warms the directory the process starts in. `--db <file>` pins the index file of
that launch root only.
The CLI takes the same value as the global `--root` flag, which defaults to the current directory.
It refuses a home directory or a filesystem root the same way the tools do. `code-kb scan` runs on
any root you give it.
#### Claude Code (without the plugin)
```bash
claude mcp add --scope user code-kb -- code-kb serve
```
#### Codex (without the plugin)
In `~/.codex/config.toml`:
```toml
[mcp_servers.code-kb]
command = "code-kb"
args = ["serve"]
```
#### Antigravity CLI (without the plugin)
```bash
agy mcp add code-kb code-kb serve
```
Global config (`~/.gemini/config/mcp_config.json`) with `"eager": true`. The same config works
for the AGY CLI and the Antigravity IDE:
```json
{
"mcpServers": {
"code-kb": {
"command": "code-kb",
"args": ["serve"],
"disabled": false,
"eager": true,
"force_all_tools_eager": true
}
}
}
```
Lifecycle hook configuration (`~/.gemini/config/hooks.json`):
```json
{
"code-kb": {
"PreInvocation": [
{
"command": "code-kb hook PreInvocation",
"timeout": 10,
"type": "command"
}
]
}
}
```
Progressive disclosure skill linking:
```bash
ln -sf /path/to/code-kb/skills/code-kb ~/.gemini/config/skills/code-kb
```
#### Grok CLI (without the plugin)
Project-level `.mcp.json`:
```json
{
"mcpServers": {
"code-kb": {
"command": "code-kb",
"args": ["serve"]
}
}
}
```
#### Cursor
In `.cursor/mcp.json` at the repository root:
```json
{
"mcpServers": {
"code-kb": {
"command": "code-kb",
"args": ["serve"]
}
}
}
```
#### OpenCode
Add to `opencode.json` in the project:
```json
{
"mcpServers": {
"code-kb": {
"command": "code-kb",
"args": ["serve"]
}
}
}
```
#### Claude Desktop
Claude Desktop has one global config. The agent passes `project_root` on each call, so one server
works for every project.
Add to `claude_desktop_config.json` (`%APPDATA%\Claude\claude_desktop_config.json` on Windows, `~/Library/Application Support/Claude/claude_desktop_config.json` on macOS):
```json
{
"mcpServers": {
"code-kb": {
"command": "code-kb",
"args": ["serve"]
}
}
}
```
#### Other GUI Apps (Windsurf, Visual Studio, VS Code)
Use the app's MCP config file and the same arguments: `["serve"]`. To pre-warm one project's
index at start, add `"--root", "/absolute/path/to/project"`. On Windows write the path with
forward slashes, for example `C:/source/project`.
#### GitHub Copilot CLI & Terminal Agents
For terminal harnesses (Copilot CLI, Pi, Swival, Zed), configure the MCP server to run
`code-kb serve`.
To remove a manual configuration, delete the `code-kb` entry from the harness config
(`claude mcp remove code-kb`, `agy mcp remove code-kb`, or edit the file) and delete the
binaries from your `PATH`.
---
## Qt and QML
For a Qt developer, `code-kb` answers the everyday questions about a QML code base.
`file_skeleton` prints the component as its object tree: declared properties, signals,
functions, inline components, and every nested object under its owner.
`find_references` crosses component files: it lists the components that extend a base
type as `extends` rows, the plain and qualified instantiations (`Kirigami.Page` as well
as `Page`), the files that read a singleton, and the signal handlers, each labelled
`handler` when the receiver names the owner and `handler (candidate)` when it does not.
`qmldir` module files and `.qmltypes` type descriptors are indexed. Qt JavaScript files
parse, including the `.pragma library` and `.import` directives. KDE test files under
`autotests/` and files named `tst_*.qml` are hidden from
search by default; `--include-tests` shows them.
Current source builds use julie-extract 3.7.0. QML references retain inline component
ownership, respect shadowed `id` names, and resolve `parent` and `this` from their
enclosing object. QML `import` directives appear as structural facts. Qt JavaScript
`.import` symbols preserve their exact spans, so body reads exclude trailing comments.
`code-kb facts pragma --path src` lists `.pragma` facts from both QML and Qt JavaScript.
A Qt C++ header indexes the same way. `file_skeleton` prints one `Q_PROPERTY(...)` row
per declared property under its class, keeping the `READ`, `WRITE`, `NOTIFY`, and
`MEMBER` accessors the macro names. A method declared in a `Q_SIGNALS:` section is an
`event` row, and the skeleton marks it `// event` before its line range because its C++
signature reads like a method. A method in a `Q_SLOTS:` section carries `qt_slot`, and a
`Q_INVOKABLE` method carries `qt_invokable`. A class records the QML element name it declares with
`QML_ELEMENT` or `QML_NAMED_ELEMENT`. A forward declaration such as `class ColumnView;`
emits no row, so `lookup_symbol` of a class name returns one row, at its definition.
`find_structural_facts` takes a `property` alias that covers the QML property
declarations and the Qt C++ `Q_PROPERTY` facts together, each fact naming its class.
Property facts expose their available Qt metadata, including `property_type`, accessors,
and the optional `designable`, `scriptable`, `stored`, `user`, and `revision` attributes.
Until upstream header updates preserve C++ language detection, one batch of changed `.h`
files triggers a content-aware workspace scan that only re-extracts changed files.
First-class Qt support: QML, `qmldir`, `.qmltypes`, Qt JavaScript, and Qt C++ headers,
validated on pinned corpus revisions; static reference results have documented limits;
`.ui` and CMake files are not indexed.
### Historical 1.6.0 / Julie 3.3.0 measurements
| Corpus | Pinned commit | QML files | C++ files |
| :--- | :--- | ---: | ---: |
| Omarchy shell (Quickshell) | `49306774` | 106 | 0 |
| KDE Kirigami | `ca7d636` | 225 | 102 |
| KDE plasma-workspace | `a45871a` | 222 | 1051 |
| Quickshell examples | `c6d1236` | 14 | 0 |
These corpus revisions and examples were measured with code-kb 1.6.0 and julie-extract
3.3.0. They are historical measurements, not 3.3.1 results.
```bash
# Omarchy
code-kb skeleton shell/Ui/Button.qml
# 71 lines for a 209-line file; the object tree nests, so ToolTip, Row, MouseArea,
# and HoverHandler hold their own children.
code-kb refs BarWidget --file shell/Ui/BarWidget.qml --limit 200
# 15 rows: 12 `extends` rows, one per component that extends BarWidget.
# The other 3 rows are signal handlers, grouped one row per file.
code-kb refs Color --file shell/Commons/Color.qml --limit 200
# 53 rows, one per file that reads the Color singleton.
code-kb refs clicked --file shell/Ui/Button.qml
# 20 rows: 4 emit sites, 1 `handler` row, and 15 `handler (candidate)` rows.
code-kb lookup SpeedDial
# 1 row: class SpeedDial [shell/Ui/SpeedTestOverlay.qml:206-411],
# signature `component SpeedDial: Item`, an inline component.
# Kirigami
code-kb refs Page --file src/controls/Page.qml --limit 200
# 60 rows, of which 7 are `extends` rows (four of them inline components).
# Kirigami, Qt C++
code-kb skeleton src/layouts/columnview.h
# 231 lines for a 785-line file, 38 `Q_PROPERTY(...)` rows, no parse-error banner.
code-kb lookup ColumnView
# 1 class row: `class ColumnView [src/layouts/columnview.h:276-783]`.
# The forward declaration in the same file adds no second row.
code-kb facts property --path src/layouts/columnview.h --limit 5
# 5 of the file's 38 `cpp.qt_property.v1` facts, each naming the class that owns it
# (`ScrollIntentionEvent`, `ColumnViewAttached`).
```
### Known limits
- A project macro that the extractor does not know is not blanked before the parse, so it
can still cause a parse diagnostic. Diagnostics remain on both C++ corpora: 31 on
Kirigami and 298 on plasma-workspace, down 97% and 94% from julie-extract 3.2.0. None
of them is macro-shaped; a file-local `#define` used as a statement accounts for 12 of
Kirigami's 31.
- `Q_DECLARE_FLAGS(Modes, Mode)` and a `Q_OBJECT_BINDABLE_PROPERTY(...)` member are
blanked as whole statements, so the typedef and the member emit no row.
- A macro used inside a line, such as `Q_ARG(bool, true)` in a `QMetaObject::invokeMethod`
call, is left alone by design, because rewriting it would delete real arguments.
- `.ui` designer files and CMake files are not indexed.
- A QML module imported through an alias resolves to a workspace file only when the alias
is not a Qt module. `QtQuick.*` and `QtQml.*` aliases never name a workspace symbol.
- A reference list caps at 200 rows. For a heavily used singleton, `refs` groups the
member accesses per file, one row per file with a count.
---
## MCP Tool Catalog
Every tool except `telemetry_summary` requires `project_root`, the absolute path of the project or
git worktree you work in. The table does not repeat it. `path` and `file_path` are relative to
`project_root`, or absolute inside it.
| Tool | Purpose | Key Parameters | Aliases |
| :--- | :--- | :--- | :--- |
| `codebase_outline` | High-level architectural orientation of directory layout & symbols. | `path` (opt), `depth` (opt, default 2) | `dir`, `subpath` |
| `file_skeleton` | File outline with function & method bodies stripped (80–90% token savings); a directory returns its outline. | `file_path` (req) | `file`, `path` |
| `lookup_symbol` | Fast identifier lookup (exact name or prefix) across repo or scoped path. Test functions, test containers, and rows from test files are hidden unless `is_test` is true, except a row whose name equals the query. | `query` (req), `path` (opt), `kind` (opt), `is_test` (opt), `limit` (opt) | `name`, `q` |
| `search_symbols` | Conceptual search over symbol names, signatures & docstrings; substrings inside identifiers are found (`sha256` finds `parseSha256Sidecar`). Results are reranked by crediting each query term once from its strongest field (name, signature, docstring), weighted by the term's rarity across the index, plus kind and path priors; `score` is that rerank score. Test functions, test containers, and rows from test files are hidden unless `is_test` is true. | `query` (req), `path` (opt), `kind` (opt), `is_test` (opt), `limit` (opt) | `name`, `q` |
| `get_symbol_body` | Slices the source of a symbol from disk as written, with its declaration and decorators; a name with overloads in one class returns each overload. | Exactly one `symbol_name` or current-index `symbol_id`; `file_path` (opt, a file or a folder) | `symbol`, `name`, `path` |
| `get_symbol_context` | Surgical bundle: target body + callee signatures + parameter types + tests; up to four overloads in one class return a bundle each. | Exactly one `symbol_name` or current-index `symbol_id`; `file_path` (opt, a file or a folder), `include_external` (opt, def: false) | `symbol`, `name`, `path` |
| `find_references` | Callers or callees of a symbol, matched by name from AST call sites and ranked by same file, same directory, then receiver type; callers also include type usages and member accesses (filters external stdlib noise; qualify overloaded names). | Exactly one `symbol_name` or current-index `symbol_id`; `file_path` (opt), `direction` ("callers" \| "callees", def: callers), `include_external` (opt, def: false) | `symbol`, `name`, `file`, `path` |
| `blast_radius` | Multi-hop reverse reachability (CTEs) & targeted test prediction. | `symbol` or current-index `symbol_id` (opt), `file` (opt guard for an ID), `depth` (opt, def: 2), `limit` (opt) | `name`, `path`, `impact` |
| `find_structural_facts` | Queries framework facts (routes, SQL queries, config keys, tables). Lists all categories when omitted. | `category` (opt), `path` (opt), `limit` (opt) | `cat`, `kind`, `type`, `file`, `file_path` |
| `telemetry_summary` | Token savings with their coverage, call counts, and error rates from `~/.code-kb/telemetry.db`, across all workspaces or scoped to the project of the most recent code-kb call. | `time_window` (opt, def: all), `workspace_only` (opt, def: false), `version` (opt, def: current; `all` for every version), `json` (opt) | `since`, `window` |
---
## CLI Commands (Direct Terminal Usage)
Every MCP capability can be executed directly from your terminal with 1:1 parity:
```bash
# Architectural outline of current repo (depth 2)
code-kb outline
# Skeleton of a specific file (implementation bodies stripped)
code-kb skeleton src/main.rs
# Lookup symbols by exact name or prefix (alias: code-kb symbol)
code-kb lookup Workspace --kind struct --path crates/code-kb-core
# Conceptual search over names, signatures, and docstrings (sha256 finds parseSha256Sidecar)
code-kb search "syntax validation concurrency"
# Same search with the rerank breakdown under each result (JSON: an `explain` object per result)
code-kb search "sha256 sidecar" --explain
# Retrieve exact implementation body of a symbol
code-kb body Workspace
# Get surgical context bundle: body + callees + types + tests (alias: code-kb slice)
code-kb context ensure_fresh_file
# Find callers or callees of a function (language-agnostically filters stdlib noise)
code-kb refs open_read_only --direction callers
# View callees including external standard library tokens
code-kb refs open_read_only --direction callees --include-external
# Predict blast radius and targeted tests to run for a symbol or file
code-kb blast-radius compute_blast_radius
# Zero-argument blast radius: auto-discovers uncommitted git working-tree changes
code-kb blast-radius
# (alias: code-kb impact)
# Query framework structural facts (or omit category to list all detected categories)
code-kb facts
code-kb facts config --path Cargo.toml
code-kb facts route --limit 10
# View active log file and recent diagnostic messages. Logs live in <project>/.code-kb/logs;
# a server started outside a project logs to ~/.code-kb/logs.
code-kb logs
# Token savings, call counts, and error rates (alias: code-kb telemetry)
code-kb stats --since month
code-kb stats --workspace
# Self-contained diagnostic bundle (versions, index facts, recent errors, log tail) with a
# pre-filled GitHub issue link; the /report-issue skill files it through `gh issue create`
code-kb bug-report --title "lookup returns nothing" --description "what went wrong"
# Output agent lifecycle hook payload (SessionStart / SubagentStart / PreInvocation)
code-kb hook SessionStart
code-kb hook SubagentStart
code-kb hook PreInvocation
```
---
## Telemetry and What "Saved" Means
`code-kb` records every tool call in `~/.code-kb/telemetry.db`. `telemetry_summary` and
`code-kb stats` report the calls, the latency, the error rate, the tokens served, and the
tokens saved.
Saved is the size, in estimated tokens, of the files the answer points into, minus the tokens
served. A skeleton, body, or context read is measured against its own file. A lookup, search,
references, blast-radius, or facts answer is measured against the distinct files its rows name,
at most 20 files. A call with no file to point at, such as an outline or a telemetry summary,
records no baseline and is not counted as saved.
The report states that coverage beside the number, so you can see how much of the window it
covers:
```text
Tokens Saved, upper bound: ~<saved> (baseline known for <K> of <M> calls; baseline = reading each file an answer names in full)
```
Each tool row carries the same pair as `~N (K/M)`. Rows written before this measurement existed
carry no baseline and are never rewritten.
---
## Development
For contributors building `code-kb` from source:
### Prerequisites
- [Rust](https://www.rust-lang.org/) (1.95+ / Edition 2024)
- Extractor binary: Run `./scripts/restore-julie-extract.sh` to download the pinned [`julie-extract`](https://github.com/anortham/julie-extractors) binary
### Build and Install Locally
```bash
# Restore pinned julie-extract binary
./scripts/restore-julie-extract.sh
# Install code-kb binary from local checkout
cargo install --path crates/code-kb-cli --force
```
### Running Tests & Verification
```bash
# Run test suite
cargo test --workspace
# Run the plugin launcher and manifest tests
node --test tests/plugin/*.test.cjs
# Run release pre-flight verification
./scripts/release-preflight.sh
```
To run the Claude Code plugin from this checkout, build and load it. The launcher runs
`target/release/code-kb` when it exists, so no download happens:
```bash
cargo build --release
claude --plugin-dir .
```
---
## Architecture & Engineering Plans
`code-kb` was designed based on extensive benchmarking and retrospectives from earlier code intelligence engines:
- [**001: Architecture & Service Model**](docs/plans/001-architecture-and-service-model.md) — Multi-project workstation scope, Git worktree deduplication, and RAG vs. AST evaluation.
- [**002: Workspace Scoping & MCP**](docs/plans/002-workspace-scoping-and-mcp.md) — 1:1 session binding, eliminating workspace registries and `workspace_id` friction.
- [**003: Tool Catalog & Schema**](docs/plans/003-tool-catalog-and-schema.md) — Historical tool-catalog decisions and token-minimized output formats.
- [**004: File Synchronization & Watchers**](docs/plans/004-file-synchronization-and-watchers.md) — 3-tier sync: tool-driven updates, JIT staleness guards, and debounced background watching.
- [**005: Cold-Start Reconciliation**](docs/plans/005-startup-reconciliation.md) — Detecting and reconciling offline edits in under 50ms on startup.
- [**006: Retrospective Lessons from Miller**](docs/plans/006-lessons-from-miller.md) — Analysis of calibration data, performance ledgers, and traps to avoid.
- [**010: Master Implementation Plan**](docs/plans/010-master-implementation-plan.md) — The phased engineering roadmap from workspace scaffolding to release.
- [**ADR 001: Zero Workspace Parameters**](docs/decisions/001-zero-workspace-parameters.md) — The old rule against workspace parameters in tool schemas. The required `project_root` parameter replaced it in 2.1.0.
- [**AGENTS.md Guidelines**](AGENTS.md) — Strict architectural invariants and rules for AI coding assistants working in this repository.
---
## License
Dual-licensed under MIT or Apache-2.0.