@lumoai/cli 1.37.0 → 1.38.0

package/assets/skill/SKILL.md

@@ -85,7 +85,7 @@ The command catalog below is a **map**: it lists every command grouped by domain
 
 **Cost (per-operation token read-out)** — see [task-context.md](references/task-context.md)
 
-- `lumo cost [--task <id>|--session <id>|--since <date>] [--by tool|model|member|session] [--json]` — per-operation (per-tool) token cost read-out. Attributes each model step's token delta to the tool(s) it ran (per-step where POST_TOOL_BATCH data exists, per-turn fallback otherwise), output vs cache_read shown separately, plus a per-step coverage line and a "heuristic" note (parallel tools split a step's tokens evenly). Scope is mutually exclusive: `--task` (one task) / `--session` (one Claude Code session) / `--since` (workspace window); default = workspace last-30-days. `--by` only changes which grouping is the headline (the others are still printed when non-trivial). For the per-task Top-5 inline, see `lumo task lineage`.
+- `lumo cost [--task <id>|--session <id>|--since <date>] [--by tool|model|member|session] [--json]` — per-operation (per-tool) token cost read-out. Attributes each model step's token delta to the tool(s) it ran (per-step where POST_TOOL_BATCH data exists, per-turn fallback otherwise); each ranking row breaks the cost into output / input / cache_create / cache_read columns plus total, with a per-step coverage line and a "heuristic" note (parallel tools split a step's tokens evenly; output and cache_read attribute cleanly per step while input and cache_create are bursty, so per-tool figures for those two are coarse). Scope is mutually exclusive: `--task` (one task) / `--session` (one Claude Code session) / `--since` (workspace window); default = workspace last-30-days. `--by` only changes which grouping is the headline (the others are still printed when non-trivial). For the per-task Top-5 inline, see `lumo task lineage`.
 
 **Artifacts & Figma** — see [artifacts-figma.md](references/artifacts-figma.md)
 

package/assets/skill/references/task-context.md

@@ -187,8 +187,11 @@ Per-operation (per-tool) token cost read-out. Where `task lineage` answers
 token delta to the tool(s) that ran in it — **per-step** where the
 `POST_TOOL_BATCH` hook captured the tool list, **per-turn fallback** otherwise
 (a parallel-tool step splits its tokens evenly across the tools, hence the
-"heuristic" note). `output` (generation) and `cache_read` (~95%, structural —
-turns × context) are shown in separate columns.
+"heuristic" note). Each ranking row breaks the cost into four columns —
+`output` (generation), `input`, `cache_create`, `cache_read` (~95%, structural —
+turns × context) — plus `total`. `output` and `cache_read` attribute cleanly
+per step; `input` and `cache_create` are bursty (prompt + cache checkpoints), so
+their per-tool figures are coarser.
 
 ```bash
 lumo cost --task LUM-42

package/dist/cli/src/commands/cost.js

@@ -10,15 +10,19 @@ function groupThousands(value) {
         .toString()
         .replace(/\B(?=(\d{3})+(?!\d))/g, ',');
 }
+/** Render one ranking row: label + the four token categories + total, aligned. */
+function rankRow(label, output, input, cacheCreate, cacheRead, total) {
+    return `  ${label.slice(0, 20).padEnd(20)}  ${output.padStart(10)}  ${input.padStart(10)}  ${cacheCreate.padStart(12)}  ${cacheRead.padStart(12)}  ${total.padStart(12)}`;
+}
 function rankTable(title, rows) {
     if (rows.length === 0)
         return '';
     const lines = [
         title,
-        '  operation              output      cache_read       total',
+        rankRow('operation', 'output', 'input', 'cache_create', 'cache_read', 'total'),
     ];
     for (const r of rows.slice(0, 20)) {
-        lines.push(`  ${r.label.slice(0, 20).padEnd(20)}  ${groupThousands(r.output).padStart(10)}  ${groupThousands(r.cacheRead).padStart(12)}  ${groupThousands(r.total).padStart(12)}`);
+        lines.push(rankRow(r.label, groupThousands(r.output), groupThousands(r.input), groupThousands(r.cacheCreation), groupThousands(r.cacheRead), groupThousands(r.total)));
     }
     return lines.join('\n');
 }
@@ -39,7 +43,7 @@ function formatOperationCost(data, by) {
         : `${Math.round(data.coverage.perStepPct * 100)}%`;
     const out = [];
     out.push(`Per-operation cost — ${data.scope.kind} ${data.scope.label}`);
-    out.push(`Total: output ${groupThousands(data.grandTotal.output)} · cache_read ${groupThousands(data.grandTotal.cacheRead)} · all ${groupThousands(data.grandTotal.total)} tokens`);
+    out.push(`Total: output ${groupThousands(data.grandTotal.output)} · input ${groupThousands(data.grandTotal.input)} · cache_create ${groupThousands(data.grandTotal.cacheCreation)} · cache_read ${groupThousands(data.grandTotal.cacheRead)} · all ${groupThousands(data.grandTotal.total)} tokens`);
     out.push(`Per-step attribution: ${pct} of ${data.coverage.toolTurns} tool-using turns (rest fall back to per-turn split)`);
     out.push('');
     out.push(rankTable(`By ${by}:`, primary));
@@ -52,7 +56,7 @@ function formatOperationCost(data, by) {
     if (by !== 'session' && data.bySession.length > 1)
         out.push('\n' + rankTable('By session:', data.bySession));
     out.push('');
-    out.push('Note: token→tool attribution is heuristic — a model step that fires several tools in parallel splits its tokens evenly across them; cache_read (~95%) is structural (turns × context), shown alongside output (generation).');
+    out.push('Note: token→tool attribution is heuristic — a model step that fires several tools in parallel splits its tokens evenly across them. output (generation) and cache_read (~95%, structural — turns × context) attribute cleanly per step; input and cache_create are bursty (prompt + cache checkpoints) and per-tool figures for those two are coarse.');
     return out.join('\n');
 }
 async function cost(opts) {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lumoai/cli",
-  "version": "1.37.0",
+  "version": "1.38.0",
   "description": "Lumo CLI — manage tasks and sessions from the terminal",
   "license": "MIT",
   "author": "cli@uselumo.ai",