effort-cycle
Reviewed
at commit 6787d3e
Alt+E and Alt+Shift+E step the effort level without a transcript row, and the footer shows the model and level as a colored meter.
0 upvotes
Log in to upvote or save- Made by
- Aner Cohen
- Pinned commit
-
6787d3eca0d9c5f3425d2a145cac9097e725091b - What we checked
-
Ctrl+Up and Ctrl+Down raise or lower the effort level of the main thread or the subagent you are viewing, shown as a colored meter in the footer and in subagent rows. It reads your Claude settings and the session's models and agents, makes no network calls, and writes small row files under ~/.claude/subagent-rows. An optional setting resizes the terminal by one column to refresh rows.
Review details
Static review of the source at the pinned commit; not run. Hooks: session.start, turn.step (sets the effort field of model requests to the level you picked), agent.spawn, command.run for /effort and /model, ui.render (footer SessionMode, AbovePrompt hidden key buttons, Agent ToolUse line), ui.message and ui.fault; footer carets are a Client module (caret.tsx) that only posts click counts. State is kept with $.state. It reads settings from every source via $.settings.read and CLAUDE_CODE_EFFORT_LEVEL to work out the starting level. It sets the env var EFFORT_CYCLE_ROWS, and its settings.json ships a subagentStatusLine command that runs python3 -I -S on subagent-rows/rows.py (every 5 s while subagents run; unset means no output). rows.py is stdlib-only: it reads JSON from stdin and ~/.claude/subagent-rows/sessions/<session>/*.json, prints row text, and deletes session folders there untouched for over a week. The mod writes effort-cycle.json into that session folder. With the nudgeRows option (off by default) it runs python3 -c on a script that narrows the terminal one column through the controlling tty and restores it 5 ms later. A subagentStatusLine in the user's own settings overrides the shipped one. No network, no model calls.
- What it uses
-
- Reads your prompts: no
- Sees tool calls: no
- Blocks or changes tool calls: no
- Reads or writes files: yes
- Runs programs on your machine: yes
- Calls the model (uses tokens): no
- Sends data over the network: no
- Stores data: yes
- Shows something on screen: yes
- Hooked events
Alt+E steps the effort level up and Alt+Shift+E down, for the main thread or the subagent in view; the footer shows that agent's model and level as they change, with a ‹ and › to click while the pointer is over it
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/effort-cycle/)