All Claude Code modsdevohmycode/claude-mods
clauget
The scale, then the diet: and the diet only where the scale says it pays. Each model request inside a turn streams through one hook, whose stop chunk carries the four token counters and the model that answered, so the journal has a line per step, the main loop's and every subagent's apart, and names the cache misses where they happen. /clauget opens the report: the window and the cost, the bill of the prefix by provider with how often each provider's tools were called, the break-even of deferring, the instruction files, and every lever with what it saved and what it cost. Out of the box the mod measures and changes nothing. /clauget on, off or strict turns the levers on from the next session: a policy on the prefix decided once and never moved mid-session, tool results cut to the shape of their fat and filed whole so one Read undoes a cut, a small-prefix explorer agent, a lighter transcript for the compaction, economy tiers that tighten with the plan window, a forecast of when it fills, and prompts deferred to its reset. Each lever was measured on a real session; the ones that lost were closed and are named in the README. Nothing the mod writes reaches the model unless a lever puts it there, and then it is counted in the mod's own column.
review UI rewrite
In its last recorded footprint, clauget reaches L2 (writes or runs) by its classified calls and passed claude plugin validate on Claude Code 2.1.300.
- Events hooked: 22
- API methods referenced: 27
- Access from classified calls: L2, writes or runs: writes files, changes config, drives Claude, reads files, reads env vars, persists state, draws
- Unclassified calls: 4 (
$.agent.register,$.cockpit.tab,$.cockpit.tabs,$.tool.list), full access unknown - Last recorded footprint, revision
dcb7818: Repository could not be scanned; showing the last known footprint. - Static scan, not a runtime test
Install hint
/plugin install clauget --marketplace devohmycode/claude-modsFrom the repository's own marketplace. Confirm the plugin name in the repository's README before installing. Mods need Claude Code 2.1.287 or later. The latest scan could not confirm this mod, so the command may be out of date.
Access
L2 Writes or runsother: $.agent.register, other: $.cockpit.tab, other: $.cockpit.tabs, other: $.tool.list, writes files, changes config, drives Claude, reads files, reads env vars, persists state, drawsValidation
not verifiedView the source on GitHub Back to the list
What the scan recorded
- Observed events
- subagent spawns, classic hooks, every tool call, every prompt, compaction, the system prompt
- Hooks
engine.create, session.start, turn.step, agent.spawn, classic.PreModelSwitch, turn.complete, tool.call, prompt.submit, session.end, tool.describe, agent.offer, command.describe, prompt.attachment, session.compact, classic.SessionStart {source=resume}, prompt.section, prompt.context, session.measure, command.run {command=clauget}, ui.invalidate, config.set, ui.render {component=Pane}- API calls
$.agent.register, $.clock.after, $.clock.every, $.clock.now, $.clock.sleep, $.cockpit.tab, $.cockpit.tabs, $.command.register, $.config.set, $.env.get, $.fs.read, $.fs.stat, $.fs.write, $.model.fork, $.prompt.submit, $.session.compact, $.session.id, $.session.usage, $.store.delete, $.store.get, $.store.keys, $.store.set, $.tool.list, $.turn.abort, $.ui.log, $.ui.resolve, $.ui.status- Validation on 2.1.300
- not verified
- Last known result
- Passed on 2.1.300
- Validator output
- Repository could not be scanned; showing the last known footprint.
- Scanned revision
dcb78189f212- Marketplace warnings
- claude-devohmycode-mods
- UI rewrite review
- UI hook source contains control-character strings. Review any values passed to next() as rewritten props.text; the scanner does not trace whether these strings reach that call. claude-clauget-mod/hooks/cuts/noise.ts:53
These are the hooks and API calls reported by claude plugin validate. Access levels describe the widest reach of those calls and are not safety ratings. Validation is a static check, not a runtime test, and a passing result does not prove that a mod works or is safe. How the scan works.
More from devohmycode/claude-mods
- cockpitOne pane of tabs beside the transcript, and the $.cockpit noun other plugins fill. /cockpit opens it, and so…
- contextmanagerWatches a session for wasted context and time, shows a one-click card when a waste behaviour repeats, and let…
- forgeMake your own plugins and mods, and catch what a mod breaks without a word before a session loads it. /forge…
- messageWhat another session sends lands in a tab instead of in the context. A peer's delivery is written to the mail…
- switchA pane of the models and the effort levels, each switched with one click. /switch opens it; /switch opus or /…
Badge for your README
Authors can show the latest scan result in their README. Each badge links to this page.
[](https://mods.aidojo.si/mods/devohmycode--claude-mods--clauget/)
[](https://mods.aidojo.si/mods/devohmycode--claude-mods--clauget/)