antigravity-plugin-context-mode 1.1.0 → 1.3.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/CHANGELOG.md CHANGED
@@ -2,6 +2,47 @@
2
2
 
3
3
  All notable changes to `antigravity-plugin-context-mode` will be documented in this file.
4
4
 
5
+ ## [1.3.0] - 2026-11-15
6
+
7
+ ### Added
8
+ - **Subagent & Primary Agent Unified Lifecycle Hooks**: `PreToolUse` now evaluates all lifecycle checks (un-paged file reads, unbounded command dumps, active cutoffs, and file category exemptions) identically across both primary agents and subagents.
9
+ - **Shared Conversation Dashboard for Subagents**: Interceptions and metrics generated by subagents now resolve to the root/parent conversation ID (`resolveSharedSessionId`), feeding directly into the active conversation dashboard in real-time. Added dedicated subagent badges in the recent activity feed to indicate which subagent triggered the event.
10
+ - **Session Lifecycle Overrides Reset**: Added `clearSessionOverrides()` to `ContextModeStore`. Wired into the `SessionStart` hook and server process initialization so that when a new application session begins (app restart), transient session overrides are cleared and the session starts fresh with the Lifetime default.
11
+ - **Direct Indexing Telemetry (`ctx_index`)**: `ctx_index` now records an event in the `events` journal upon successful indexing, immediately updating dashboard call counts and crediting kept-out tokens.
12
+
13
+ ### Fixed
14
+ - **Lifetime Cutoff Propagation to New Sessions and Conversations**: Changing the cutoff in Lifetime scope (`"all"`, `"lifetime"`, or `null`) now updates `cutoff:global` in SQLite and clears any transient session-level (`cutoff:session:app`) and fallback default (`cutoff:session:default`) overrides. New sessions and new conversations automatically inherit the new Lifetime cutoff as their default.
15
+ - **Accurate Context Saved Calculations (Eliminated Double-Counting)**: Refactored `getStats()` to prevent execution outputs from being counted twice across `events` (`sandbox_bytes`) and `sources` (`bytes_total`). Raw, returned, and kept-out bytes are now computed on a per-event basis with `MAX(sandbox_bytes, bytes, returned_bytes)`, ensuring passthrough commands with full returns do not penalize the savings of other tools.
16
+ - **MCP Server Root Session Binding**: `server.js` and `createTools` now bind session ID resolution to `process.env.ANTIGRAVITY_ROOT_CONVERSATION_ID` and `process.env.ANTIGRAVITY_PARENT_CONVERSATION_ID`, ensuring MCP tool calls from subagents share indexed chunks and search states with the primary conversation.
17
+
18
+ ## [1.2.1] - 2026-11-15
19
+
20
+ ### Fixed
21
+ - **CI Test Database Contention (`CONTEXT_MODE_DB_PATH`)**: Updated `getDefaultDbPath()` in `lib/store.js` to respect `process.env.CONTEXT_MODE_DB_PATH`, and configured `tests/test_mcp.js` and `tests/test_hooks.js` to run in dedicated isolated temporary directories. Prevents database lock contention (`SQLITE_BUSY`) during concurrent test execution on multi-core CI runners.
22
+ - **SQLite Busy Timeout**: Added `PRAGMA busy_timeout = 5000;` to database initialization with fallback handling to ensure robust concurrency when multiple processes interact with SQLite in WAL mode.
23
+ - **Category Exemption Group Toggles**: Replaced the long list of individual file exemption chips in the dashboard with clean, one-click category pills (`Images`, `Documents`, `3D Models`, `Music`, `Video`, `Archives`, and `Source Code`). All binary categories default to enabled, with Source Code disabled by default. Added optimistic UI state switching in `toggleCategory()` for instant feedback upon click.
24
+ - **Streamlined Custom File Types**: Restricted the dashboard's exemption tags container to only display user-added custom file types (e.g. `.csv`, `.parquet`, `.log`) with quick removal buttons (`×`).
25
+ - **Status Banner Cleanup**: Replaced `"Live Telemetry"` with **`"Showing stats for: <SCOPE>"`** and removed distracting animations (`animate-ping`, `animate-pulse`) for steady, non-blinking visual indicators.
26
+ - **Dynamic Upper-Right Version Badge**: Displays the active plugin version badge (`v1.2.1`) in the top-right header cluster of the dashboard, populated dynamically from `package.json` via `/api/stats`.
27
+ - **Documentation & Formatting Polishing**: Removed all LaTeX math syntax from `README.md`, replacing it with standard Unicode (`≤ 50 lines`, `≤ 200 lines`, `≤ 800 lines`), and consolidated redundant configuration tables.
28
+
29
+ ## [1.2.0] - 2026-11-15
30
+
31
+ ### Fixed
32
+ - **Quoted Path Stripping in `PreToolUse`**: Added path argument sanitization to strip enclosing quotes (`"..."`, `'...'`), decode URI schemes (`file:///`), and trim line anchor fragments (`#L...`), preventing `fs.existsSync` failures on Windows when tool calls pass quoted file paths.
33
+ - **Interception Journaling & Telemetry**: Added event journaling on all blocked `read_file` and `run_command` tool calls directly into SQLite (`events` table). Interceptions now immediately update `tokensPreserved`, `keptOutBytes`, `reductionRate`, and the Prompt Payload Comparison bar in the sidecar dashboard.
34
+ - **Dynamic Command Inspection (No Hardcoded Extension Whitelist)**: Refactored `run_command` in `PreToolUse` (`inspectCommand`) to eliminate hardcoded extension regexes completely. Commands invoking `cat`, `type`, `Get-Content`, or `gc` now dynamically extract target paths, check them against the active SQLite cutoff, and respect user-configured exemptions. Safe small files ($\le$ cutoff) and bounded commands (`-Head`, `| Select-Object -First`) are allowed, while wildcard and unpaged dumps are blocked.
35
+ - **UTF-16 & Known Text Exemption Safeguard**: Added `KNOWN_TEXT_EXTENSIONS` check and UTF-16/UTF-8 BOM header inspection to prevent text and code files (such as `.cs` and `.md`) containing null bytes from being misidentified as binary files and bypassing size cutoffs.
36
+ - **Live Interception Feed in Dashboard**: Updated the "Recent Indexed Sources & Interceptions" dashboard panel to display blocked calls (`🛡️ Blocked <tool>: <path>`) with exact saved token metrics alongside indexed FTS5 sources.
37
+ - **Renamed 'Sandbox Cap' to 'Preview Limit'**: Replaced the confusing "Sandbox Cap" label with **`PREVIEW LIMIT`**, clearly indicating the maximum preview slice of command and script output returned into the chat.
38
+ - **Clickable 'Active' / 'Paused' Toggle**: Turned the status badge into an interactive toggle button. Users can click to switch Context Mode between **Active** (pulsing green dot) and **Paused** (amber dot). When paused, the `PreToolUse` hook allows all tool calls through without size checks or command interception, persisting the paused state across active scopes in SQLite.
39
+ - **Fixed Feed Width Truncation**: Removed the hardcoded `max-w-[210px]` container constraint and arbitrary string slice in the dashboard feed, replacing it with responsive `min-w-0 flex-1` flex sizing and clean path formatting (`formatItemPath`) so file and command paths scale dynamically to the full width of the dashboard panel.
40
+ - **Reformatted Metric Rows & Sidebar Layout**: Redesigned the 4 performance metrics into sleek, compact horizontal key-value rows inside a single card (labels on the left, values right-aligned) with subtle hairline dividers. This replaces the tall vertical 12-line card stack in normal sidebars with an information-dense, space-efficient ~140px card that naturally scales to 2 columns on wide screens without wrapping or wasting space.
41
+ - **Unified Lifecycle Gates & System Health (Moved to Bottom)**: Combined `Active Lifecycle Gates` and `Engine & Storage Health` into a single cohesive reference card with an inline `HEALTHY` badge, streamlined gate checkmarks, and compact engine specs. Moved this card to the very bottom of the dashboard so actively changing interactive panels (metrics, exemptions, search, and the live interception feed) remain prominently visible at the top.
42
+ - **Dual-Color Single Progress Bar**: Consolidated the separate raw dump and sandboxed slice progress bars into a single, compact dual-color progress bar placed directly beneath the Live Telemetry status banner. It visually depicts % of context saved in emerald green (`#00e599`) and prompt payload in rose (`#f43f5e`), with live `% saved` text. Removed the redundant middle comparison card and sub-labels since the metric cards directly below inform the user of all exact figures.
43
+ - **User-Friendly Dashboard Terminology**: Replaced cryptic architecture jargon (such as `"SQLite FTS5 Reciprocal Rank Fusion & Sandbox Execution"`, `"Porter BM25 + Trigram RRF"`, and `"Tokens diverted from LLM"`) with clear, value-focused explanations highlighting token savings, context reduction, and keeping LLMs smart, focused, and less forgetful.
44
+ - **Context-Focused Language & Removal of Log Fixation**: Changed card sub-labels from confusing "prompt bloat" phrasing to **`"Context reduced by <size>"`** and **`"Context window preserved"`**, clarifying that user prompts are never deleted or truncated. Removed all narrow fixations on "logs" throughout tooltips, input placeholders, and lifecycle gates to accurately represent all large files, code, and terminal outputs.
45
+
5
46
  ## [1.1.0] - 2026-11-14
6
47
 
7
48
  ### Added
package/README.md CHANGED
@@ -1,45 +1,46 @@
1
1
  # Antigravity Context Mode
2
2
 
3
- > **Stop context bloat. Save 90%+ of your prompt tokens. Keep your agent fast, focused, and responsive.**
3
+ > **Stop context bloat. Save 90%+ of your prompt tokens. Keep your agent fast, focused, and less forgetful.**
4
4
 
5
5
  [![npm version](https://img.shields.io/npm/v/antigravity-plugin-context-mode.svg)](https://www.npmjs.com/package/antigravity-plugin-context-mode)
6
6
  [![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)
7
7
  [![Zero Runtime Dependencies](https://img.shields.io/badge/dependencies-0-brightgreen.svg)]()
8
8
  [![Antigravity App Compatible](https://img.shields.io/badge/Antigravity%20App-Native%20Plugin-6366f1.svg)]()
9
9
 
10
- **Antigravity Context Mode** is a free, local-first plugin built specifically for the **Google Antigravity App**. It prevents terminal commands, build logs, and multi-megabyte files from flooding your conversation history, saving thousands of tokens per turn and stopping "agent amnesia" in its tracks.
10
+ **Antigravity Context Mode** is a free, local-first plugin built specifically for the **Google Antigravity App**. It prevents terminal commands, build outputs, multi-megabyte source files, and massive datasets from flooding your conversation history — saving thousands of tokens per turn and stopping "agent amnesia" in its tracks.
11
11
 
12
12
  ---
13
- <img width="962" height="812" alt="image" src="https://github.com/user-attachments/assets/873f818d-3bd9-4746-94a9-fc182a174efb" />
13
+ <img width="817" height="690" alt="Antigravity Context Mode Dashboard" src="https://github.com/user-attachments/assets/dbf6d666-58d0-48f3-bc02-9783acf3fb37" />
14
14
 
15
- ## ⚡ The Problem: Context Window Bloat
15
+ ## The Problem: Context Window Bloat
16
16
 
17
17
  When an AI coding agent runs tools, standard workflows dump everything directly into your conversation history:
18
18
 
19
19
  * Running `Get-Content Logs/Editor.log` dumps **65 KB (~16,000 tokens)** into your prompt.
20
- * Running test suites or compiler logs dumps **100+ KB** of repetitive output.
20
+ * Running test suites, compiler outputs, or git diffs dumps **100+ KB** of repetitive text.
21
21
  * Reading large source files or documentation dumps **thousands of lines** of irrelevant boilerplate.
22
22
 
23
23
  ### Why this hurts:
24
- 1. **Agent Amnesia**: Important system instructions and recent conversation steps get drowned out in a sea of raw log text ("lost in the middle").
25
- 2. **Slow Responses**: Prompt evaluation times climb rapidly as token count grows.
26
- 3. **Burned Quota & Limits**: You hit model context ceilings and rate limits after only a few turns.
24
+ 1. **Agent Amnesia**: Important system instructions, user constraints, and earlier architectural decisions get drowned out ("lost in the middle").
25
+ 2. **Slower Responses**: Model evaluation and generation times climb as the context window fills up.
26
+ 3. **Burned Quota & Ceilings**: You hit model context limits and rate limits after only a few turns.
27
27
 
28
28
  ---
29
29
 
30
- ## 🛡️ The Solution: How Context Mode Works
30
+ ## The Solution: How Context Mode Works
31
31
 
32
32
  ```text
33
33
  ┌────────────────────────────────────────────────────────────────────────┐
34
34
  │ ANTIGRAVITY AGENT │
35
35
  └───────────────────────────────────┬────────────────────────────────────┘
36
36
  │
37
- Tool Call (e.g. read large log)
37
+ Tool Call (e.g. read large file/command)
38
38
  ▼
39
39
  ┌────────────────────────────────────────────────────────────────────────┐
40
40
  │ PRE-TOOL-USE LIFECYCLE HOOK GATE │
41
- │ • Blocks unbounded dumps (cat *.log, Get-Content) │
41
+ │ • Blocks unbounded dumps (cat, Get-Content, git diff) │
42
42
  │ • Enforces 8 KB cutoff for un-paged file reads │
43
+ │ • Automatically exempts binary media, documents, & prompt attachments │
43
44
  └───────────────────┬────────────────────────────────┬───────────────────┘
44
45
  │ │
45
46
  Under 8 KB (Safe) Over 8 KB (Blocked)
@@ -54,20 +55,20 @@ When an AI coding agent runs tools, standard workflows dump everything directly
54
55
  └───────────────────────────────────┘
55
56
  │
56
57
  ▼
57
- Returns only the 550 B
58
- filtered answer to prompt!
58
+ Returns only the filtered,
59
+ relevant answer to the chat!
59
60
  ```
60
61
 
61
- 1. **Smart Lifecycle Gating (`hooks.json`)**: Before a command or file read executes, Context Mode checks if it will cause a token flood. If it exceeds 8 KB or attempts an unbounded log dump, the hook intercepts it and guides the agent to query or sandbox the data instead.
62
- 2. **"Thinking in Code" (`ctx_execute`)**: Instead of reading a 5,000-line log into memory, the agent executes a 5-line local script (Node, Python, PowerShell, or Bash). The script filters the log in **50 ms** and returns *only* the specific errors to the conversation.
63
- 3. **SQLite FTS5 Storage (`ctx_search`, `ctx_index`)**: Full raw outputs are saved safely on your local disk in `~/.context-mode/context.db`. The agent uses reciprocal rank fusion search to fetch only the exact snippets it needs.
64
- 4. **Live Sidecar Dashboard**: A native side pane right inside the Antigravity App that displays your real-time token savings and lets you search indexed logs with a click.
62
+ 1. **Smart Lifecycle Gating (`hooks.json`)**: Before a command or file read executes, Context Mode checks if it will cause a context flood. Applied identically to primary agents and subagents, the hook intercepts un-paged reads exceeding the active cutoff (default 8 KB) and unbounded dumps, guiding agents to query or sandbox the data instead.
63
+ 2. **"Thinking in Code" (`ctx_execute`)**: Instead of reading a 5,000-line log into memory, the agent executes a concise local script (Node.js, Python, or PowerShell). The script filters the data in **50 ms** and returns *only* the specific answer or error to the conversation.
64
+ 3. **Local SQLite FTS5 Storage (`ctx_search`, `ctx_index`)**: Raw outputs and indexed files are saved safely on your local disk in `~/.context-mode/context.db`. The agent uses reciprocal rank fusion search to fetch only the exact snippets it needs.
65
+ 4. **Live Sidecar Dashboard**: A native side pane right inside the Antigravity App that displays real-time context savings across primary agents and subagents, lets you toggle file categories on the fly, and search indexed data with a click.
65
66
 
66
67
  ---
67
68
 
68
- ## 📊 Proven Results
69
+ ## Proven Results
69
70
 
70
- Tested live on real Unity 6.6 build and engine logs in the Antigravity App:
71
+ Tested live on real-world build logs, source files, and datasets in the Antigravity App:
71
72
 
72
73
  | Operation | Without Context Mode | With Context Mode | Savings |
73
74
  | :--- | :---: | :---: | :---: |
@@ -77,7 +78,7 @@ Tested live on real Unity 6.6 build and engine logs in the Antigravity App:
77
78
 
78
79
  ---
79
80
 
80
- ## 🚀 Quick Start & Installation
81
+ ## Quick Start & Installation
81
82
 
82
83
  ### Option 1: Global User Install (Recommended)
83
84
  Installs Context Mode for all your projects and conversations:
@@ -105,26 +106,36 @@ npm install -g antigravity-plugin-context-mode
105
106
 
106
107
  ---
107
108
 
108
- ## 🖥️ Live Auxiliary Sidecar Dashboard
109
+ ## Live Auxiliary Sidecar Dashboard
109
110
 
110
111
  Context Mode includes a native web dashboard that opens directly in Antigravity's **Auxiliary (Side) Pane**:
111
112
 
112
113
  * **One-Click Open Pill:** Click `[Open Context Mode Dashboard](sidecar://antigravity-plugin-context-mode/dashboard/)` in chat anytime.
114
+ * **Active / Paused Toggle:** Click the status pill in the top header to instantly pause protection when you intentionally need un-gated operations, and click again to re-activate.
115
+ * **Dynamic Version Badge:** Automatically reflects the active plugin version in the top-right header.
116
+ * **Context Saved Progress Bar:** Visual indicator displaying the exact percentage of context window preserved.
113
117
  * **3-Way Scope Toggle:** Seamlessly inspect telemetry at three distinct levels:
114
- * **Conversation**: Telemetry strictly for the active conversation; clean slate on new chats.
115
- * **Session**: Metrics aggregated across all conversations since the app restarted (Language Server boot).
116
- * **Lifetime**: All-time cumulative database metrics and persistent totals.
117
- * **Interactive Cutoff Dropdown:** Switch size cutoffs on the fly (`2 KB`, `4 KB`, `8 KB`, `16 KB`, `32 KB`, `64 KB`, or `none`). Line slice limits scale dynamically (e.g. 2 KB: $\le$50 lines, 8 KB: $\le$200 lines, 32 KB: $\le$800 lines).
118
- * **Clickable Recent Sources:** Click any indexed source in the recent sources list to immediately inspect all of its indexed chunks right inside the Search Inspector.
119
- * **Exempt File Extensions Panel:** Live tag editor to add or remove custom file extensions (e.g. `.csv`, `.log`, `.json`) that bypass size limits.
120
- * **Default Exemptions & Attachments:** All chat prompt attachments (`@` mentions and drag-and-drops), PDFs, Microsoft Office documents (`.docx`, `.xlsx`, `.pptx`), and media files automatically bypass size restrictions.
121
- * **Dynamic Sandbox Output Cap:** Automatically scales script execution preview size to double the active cutoff.
122
- * **Interactive FTS5 Search Inspector:** Search your indexed logs and files directly from the UI to see what the agent has indexed.
118
+ * **Conversation**: Metrics strictly for the active chat; clean slate on new conversations. Subagent actions within the conversation feed directly into this view.
119
+ * **Session**: Metrics aggregated across all conversations since the app restarted. Automatically resets transient overrides on app reboot.
120
+ * **Lifetime**: All-time cumulative database metrics and persistent totals. Changing the cutoff in Lifetime updates the baseline defaults for all new sessions and new conversations.
121
+ * **Shared Subagent Feed:** Interceptions triggered by subagents (`researcher`, `browser`, etc.) are badged by role and displayed alongside primary agent events in the same conversation feed.
122
+ * **File Category Toggles:** Clean, one-click pills for grouping exemptions without reading massive lists:
123
+ * **Images** (`.png`, `.jpg`, `.jpeg`, `.gif`, `.webp`, `.svg`, `.bmp`, `.ico`, `.tiff`, `.psd`) — *Enabled by default*
124
+ * **Documents** (`.pdf`, `.docx`, `.doc`, `.xlsx`, `.xls`, `.pptx`, `.ppt`, `.odt`, `.rtf`) — *Enabled by default*
125
+ * **3D Models** (`.blend`, `.fbx`, `.obj`, `.gltf`, `.glb`, `.stl`, `.dae`, `.3ds`, `.step`, `.iges`) — *Enabled by default*
126
+ * **Music** (`.mp3`, `.wav`, `.ogg`, `.flac`, `.aac`, `.m4a`, `.wma`, `.midi`) — *Enabled by default*
127
+ * **Video** (`.mp4`, `.mov`, `.avi`, `.mkv`, `.webm`, `.flv`, `.wmv`, `.m4v`) — *Enabled by default*
128
+ * **Archives** (`.zip`, `.tar`, `.gz`, `.7z`, `.rar`, `.wasm`) — *Enabled by default*
129
+ * **Source Code** (`.js`, `.ts`, `.py`, `.rs`, `.go`, `.java`, `.c`, `.cpp`, `.cs`, `.html`, `.css`, etc.) — *Disabled by default to protect context*
130
+ * **Custom File Extensions:** Dedicated section displaying only user-added file extensions (e.g. `.csv`, `.parquet`, `.log`) with instant removal.
131
+ * **Chat Attachment Exemption:** Files attached directly in chat messages (`@` mentions and drag-and-drops) always bypass size cutoffs so the agent can inspect them in full.
132
+ * **Interactive Cutoff Dropdown:** Switch size cutoffs on the fly (`2 KB`, `4 KB`, `8 KB`, `16 KB`, `32 KB`, `64 KB`, or `none`). Line slice limits scale dynamically (2 KB: ≤ 50 lines, 8 KB: ≤ 200 lines, 32 KB: ≤ 800 lines).
133
+ * **Interactive FTS5 Search Inspector:** Search your indexed logs and files directly from the UI to inspect what the agent has indexed.
123
134
  * **1-Click Session Purge:** Wipe session indexes and reset counters whenever you start a fresh task.
124
135
 
125
136
  ---
126
137
 
127
- ## 🧰 Available MCP Tools
138
+ ## Available MCP Tools
128
139
 
129
140
  Once installed, the following 8 tools are registered in your Antigravity App:
130
141
 
@@ -134,14 +145,14 @@ Once installed, the following 8 tools are registered in your Antigravity App:
134
145
  | **`ctx_search`** | Searches SQLite FTS5 using BM25 stemming and Trigram matching. | Finding error codes, method names, or stack traces in past logs. |
135
146
  | **`ctx_index`** | Indexes local files, web URLs, or text into searchable chunks. | Indexing a 1,000-line documentation file or architecture spec. |
136
147
  | **`ctx_batch`** | Runs multiple shell commands in batch and indexes the combined output. | Running `git status`, `git log`, and `npm test` in one compact call. |
137
- | **`ctx_stats`** | Displays a formatted ASCII savings report and token metrics. | Checking how many tokens were saved during your session. |
148
+ | **`ctx_stats`** | Displays a formatted savings report and token metrics. | Checking how many tokens were saved during your session. |
138
149
  | **`ctx_resume`** | Returns recent indexed sources and state pointers. | Re-orienting the agent after context compaction. |
139
150
  | **`ctx_doctor`** | Checks SQLite database health, WAL mode, and active runtimes. | Verifying database integrity and platform runtimes. |
140
151
  | **`ctx_purge`** | Clears stored sources and events from SQLite. | Resetting session or worktree indexes. |
141
152
 
142
153
  ---
143
154
 
144
- ## ⚙️ Configuration & Environment Variables
155
+ ## Configuration & Environment Variables
145
156
 
146
157
  You can customize Context Mode's cutoffs at any time using the dashboard dropdown or environment variables:
147
158
 
@@ -152,11 +163,6 @@ You can customize Context Mode's cutoffs at any time using the dashboard dropdow
152
163
  | `CONTEXT_MODE_MAX_OUTPUT_BYTES` | *Dynamic* (or `16384`) | Maximum raw preview bytes returned by `ctx_execute` (defaults to 2x cutoff). |
153
164
  | `CONTEXT_MODE_TIMEOUT_MS` | `30000` (30s) | Execution timeout for sandboxed code snippets. |
154
165
 
155
- | Variable | Default | Description |
156
- | :--- | :---: | :--- |
157
- | `CONTEXT_MODE_MAX_READ_BYTES` | `8192` (8 KB) | Maximum file size allowed for an un-paged `read_file` before gating kicks in. Set to `4096` for strict mode or `16384` for relaxed mode. |
158
- | `CONTEXT_MODE_MAX_LINE_RANGE` | `200` lines | Maximum line slice allowed without pagination rejection. |
159
-
160
166
  ### How to set on Windows (PowerShell):
161
167
  ```powershell
162
168
  $env:CONTEXT_MODE_MAX_READ_BYTES = "4096" # Enforce strict 4 KB cutoff
@@ -169,7 +175,7 @@ export CONTEXT_MODE_MAX_READ_BYTES=4096
169
175
 
170
176
  ---
171
177
 
172
- ## 🔒 Privacy & Zero Dependencies
178
+ ## Privacy & Zero Dependencies
173
179
 
174
180
  * **100% Local**: All indexed data lives on your machine in `~/.context-mode/context.db`. Zero network calls are made.
175
181
  * **Zero 3rd-Party Dependencies**: Built exclusively on native Node.js (`node:sqlite` / `DatabaseSync`, `child_process`, and `http`). No native C++ builds, no `node-gyp`, and no dependency security vulnerabilities.
@@ -177,7 +183,7 @@ export CONTEXT_MODE_MAX_READ_BYTES=4096
177
183
 
178
184
  ---
179
185
 
180
- ## 🧪 Running Tests
186
+ ## Running Tests
181
187
 
182
188
  Context Mode ships with a complete test suite using Node's native test runner:
183
189
 
@@ -189,11 +195,16 @@ Verifies:
189
195
  * SQLite FTS5 schema initialization, chunking, Porter BM25, and Trigram search.
190
196
  * Reciprocal Rank Fusion (RRF) ranking.
191
197
  * Sandbox execution and process timeout termination.
192
- * `PreToolUse` and `SessionStart` lifecycle hook gating.
198
+ * `PreToolUse` and `SessionStart` lifecycle hook gating for primary agents and subagents.
199
+ * Shared conversation dashboard linking and subagent interception event journal.
200
+ * Lifetime cutoff inheritance across new sessions and new conversations.
201
+ * Mathematically accurate context savings calculations without double-counting.
202
+ * File category toggles and custom extension management.
203
+ * Active / Paused state toggle and bypass logic.
193
204
  * MCP stdio JSON-RPC protocol compliance.
194
205
 
195
206
  ---
196
207
 
197
- ## 📄 License
208
+ ## License
198
209
 
199
210
  MIT © [Phillip Webber](https://github.com/Domush)