lot-map
Reviewed
at commit f1c2757
New version, not yet reviewed
Draws a live map beside the chat where each tool call, model request and hook sends a pulse to one of ten stations, with a ring that fills as the context…
0 upvotes
Log in to upvote or save- Source
- github.com/alexeyw/lot-map
- Made by
- Aleksei Volodin
- Pinned commit
-
f1c27572672cde6046fffcf62a98f9ea488186f2 - What we checked
-
Draws a live map in a pane beside the chat: memory, files, shell, web, model, agents, skills and hooks light up as tool calls happen. It keeps only short labels (last two path parts, web hosts, tool names) and remembers the last model name per folder. It makes no network calls, reads no files and runs no programs.
Review details
Static review of the source at the pinned commit; not run. Registers session.start (registers /lot-map, opens the pane, reads the stored last model name for the folder), command.run for /lot-map, turn.start, turn.step, turn.complete, tool.call, classic.PreToolUse, prompt.submit, classic.UserPromptSubmit, classic.SessionStart, skill.prompt, session.compact and ui.render. Every handler passes the event on unchanged; the hooks only observe and draw. The only tool-call decision it touches is a settings hook's deny/ask, which it reads and hands on as it was. Runtime calls: $.store.get/set (one value per folder: the name of the last model used, stored under the folder path), $.clock.now, $.session.usage (context percent), $.command.register, $.ui.open/resolve, plugin state atom (session rows and counters, capped at 80 rows). Labels are built by classify in hooks/map.ts: file tools show the last two path parts, WebFetch and browser tools show only the host, Bash/PowerShell show the command's description or a category word (never the command line), memory MCP calls show the tool name and a short id, subagents show their description. hookReasons (first line of a hook refusal, 72 chars) and hookRecords (record numbers found in hook context) are off by default. Prompt text is only matched against a task-notification pattern; nothing of it is kept. Labels are stripped of control and direction characters, and the SVG text is XML-escaped. map.ts and text.ts contain no runtime ($) calls. The pane renders an SVG source string produced locally. PRIVACY.md in the repo agrees with the code. All of register.tsx, map.ts classify/hostOf/clean/drawMap entry points and text.ts were read; the long drawing geometry in map.ts was skimmed, with targeted search for runtime calls, network and file access (none found). No npm dependencies.
- What it uses
-
- Reads your prompts: no
- Sees tool calls: yes
- Blocks or changes tool calls: no
- Reads or writes files: no
- Runs programs on your machine: no
- Calls the model (uses tokens): no
- Sends data over the network: no
- Stores data: yes
- Shows something on screen: yes
A live map beside the chat: memory, files, shell, web, model, agents, skills and hooks light up as they are used
Made this mod? Add a badge to your README
Paste this into your README. The badge says reviewed once the current version has been reviewed, and links readers to this page.
[](https://modsforclaude.com/mods/lot-map/)