From octo
Displays a cost breakdown by provider and workflow for the current session, with cumulative history and estimated costs.
How this command is triggered — by the user, by Claude, or both
Slash command
/octo:costsThis command is limited to the following tools:
The summary Claude sees in its command listing — used to decide when to auto-load this command
# Cost Dashboard (/octo:costs) **Your first output line MUST be:** `🐙 Octopus Cost Dashboard` Display a cost breakdown by provider and workflow for the current session (and cumulative history). ## EXECUTION CONTRACT (Mandatory) When the user invokes `/octo:costs`, you MUST follow these steps in order. ### STEP 1: Locate Usage Data Search for session usage data in these locations (check all, use whichever exist): Use the Bash tool to list and read files: ### STEP 2: Parse Provider Usage For each provider found in the usage data, extract: - **Tokens in** (input/prompt tokens) - ...
Your first output line MUST be: 🐙 Octopus Cost Dashboard
Display a cost breakdown by provider and workflow for the current session (and cumulative history).
When the user invokes /octo:costs, you MUST follow these steps in order.
Search for session usage data in these locations (check all, use whichever exist):
~/.claude-octopus/usage/ # Per-session usage logs
~/.claude-octopus/routing.log # Routing decisions with timestamps
~/.claude-octopus/sessions/ # Session state files
.claude-octopus/ # Project-local usage data
Use the Bash tool to list and read files:
ls -la ~/.claude-octopus/usage/ 2>/dev/null || echo "No usage directory"
ls -la ~/.claude-octopus/routing.log 2>/dev/null || echo "No routing log"
ls -la ~/.claude-octopus/sessions/ 2>/dev/null || echo "No sessions directory"
ls -la .claude-octopus/ 2>/dev/null || echo "No project-local usage data"
For each provider found in the usage data, extract:
Format as a clean ASCII table:
Provider Cost Breakdown
============================================================
Provider Tokens In Tokens Out Queries Est Cost
------------------------------------------------------------
Claude Opus 4.6 45,200 12,800 3 $0.55
Claude Sonnet 4.6 128,000 34,500 12 $0.24
Codex CLI - - 8 $0.64
Gemini CLI - - 4 $0.08
Perplexity - - 2 $0.06
------------------------------------------------------------
TOTAL 29 $1.57
============================================================
For providers where only query counts are available (Codex, Gemini, Antigravity, Perplexity), use the midpoint or included-access estimate from the reference table. Show $0.00 for free providers or unused providers.
Group costs by workflow/command that triggered them:
Workflow Cost Breakdown
============================================================
Workflow Providers Used Queries Est Cost
------------------------------------------------------------
/octo:discover Claude, Codex, Gemini, agy 8 $0.42
/octo:develop Claude, Codex 6 $0.35
/octo:review Claude, Codex, Gemini, agy 9 $0.58
/octo:debate Claude, Codex, Gemini, agy 6 $0.22
------------------------------------------------------------
TOTAL 29 $1.57
============================================================
Show both the current session totals and cumulative totals (if historical data exists):
Session Summary
============================================================
Current Session: $1.57 (29 queries, started 2h 15m ago)
Cumulative (7d): $8.42 (156 queries across 12 sessions)
Cumulative (30d): $34.18 (612 queries across 47 sessions)
============================================================
If cumulative data is not available, show only the current session.
If no usage data exists at all, display:
No usage data found.
Claude Octopus tracks provider usage in:
~/.claude-octopus/usage/ (per-session logs)
~/.claude-octopus/routing.log (routing decisions)
Usage data is recorded automatically when you run workflows like:
/octo:discover - Multi-AI research
/octo:develop - Multi-AI implementation
/octo:review - Multi-AI code review
/octo:debate - Multi-AI deliberation
/octo:embrace - Full 4-phase lifecycle
Run any multi-AI workflow to start tracking costs.
These are the current per-provider cost estimates used for calculations:
| Provider | Input | Output | Per-Query Estimate |
|---|---|---|---|
| Claude Opus 4.6 | $5/MTok | $25/MTok | varies by tokens |
| Claude Sonnet 4.6 | $0.80/MTok | $4/MTok | varies by tokens |
| Codex CLI | - | - | ~$0.01-0.15/query |
| Gemini CLI | - | - | ~$0.01-0.03/query |
Antigravity CLI (agy) | - | - | Included with user's Antigravity access/subscription |
| Perplexity | - | - | ~$0.01-0.05/query |
Notes:
claude-opus agent type/octo:costs # Show current session costs
/octo:costs # After running several workflows
npx claudepluginhub nyldn/claude-octopus --plugin octo/cost-captureGuides through AI provider cost dashboards, captures spend and token usage, compares to previous snapshot, and updates MODEL_ROUTING.md with routing suggestions.
/cost-trackerDisplays an estimate of current session costs, breaks down top cost drivers, and provides optimization tips based on task type. Also shows budget guidance for common task categories.
/claude-pulseDisplays real-time session dashboard with cost analysis, token breakdown, tool performance, context usage, and optimization tips. Also supports configure, stats, setup subcommands.
/budgetTracks and reports AI model usage costs from a local JSON log. Shows recent tasks with tier, model, tokens, and cost. Also supports --monthly and --reset flags.
/cc-perfAnalyzes Claude Code session token usage, costs, inefficiencies, and tool calls; provides performance summaries, optimization recommendations, model comparisons, and budget tracking.
/statsDisplays Claude Code spending analytics for today, week, month, all, or filtered by tag or branch. Defaults to week when no arguments are given.