Context Budget Manager

Manage the token budget for Claude Code sessions. Model-aware — picks the right effective context window and prices per model. The session's REAL model is read from the transcript by the budget hook; config.model is only the fallback. Everything from Sonnet 4.6 up is 1M; Haiku 4.5 is 200K.

Parse $ARGUMENTS:

status (or no arguments)

Show current budget config + auto-compact settings: Read ~/.claude-context-optimizer/config.json and ~/.claude-context-optimizer/budget-config.json (either may be missing). If no config exists, show defaults (200K working budget on a 1M window, opus-5.5 fallback model, warn at 50/70/85/95%).

set <tokens>

Update the budget limit. Parse the token count (200K, 1M, 500000 all OK). Update ~/.claude-context-optimizer/config.json (the user approves the write):

{
  "budgetTokens": <parsed_number>,
  "warnAt": [50, 70, 85, 95],
  "autoCompactAt": 90,
  "model": "opus-5.5"
}

If budgetTokens exceeds the chosen model's context window, warn the user.

model <name>

Set the FALLBACK model for cost estimation (used only when a session's transcript has no model id yet). Supported keys:

• haiku-4.5 (alias haiku) — $1/$5 per M, 200K
• sonnet-4.6 (alias sonnet) — $3/$15 per M, 1M
• sonnet-5 — $2/$10 per M, 1M
• sonnet-5.5 — $2/$10 per M, 1M
• opus-4.7 / opus-4.8 — $5/$25 per M, 1M
• opus-5 (alias opus) — $5/$25 per M, 1M
• opus-5.5 (default) — $4/$20 per M, 1M, cache reads at 0.05×
• fable-5 — $10/$50 per M, 1M
• fable-5.1 (alias fable) — $10/$50 per M, 1M, cache reads at 0.025× (vs 0.1× elsewhere)

opus-4.8-1m / opus-4.7-1m / opus-extended are back-compat aliases only — there is no 1M surcharge; the 1M window is standard at $5/$25.

Update the model field in config.json. When switching to a 1M-context model and the current budgetTokens is below 500K, ask if the user wants to bump it to 1M.

auto <on|off>

Toggle auto-compact at thresholds (80% / 90%). Update ~/.claude-context-optimizer/budget-config.json:

• auto on → autoCompactEnabled: true
• auto off → autoCompactEnabled: false

Defaults if file missing:

{
  "autoCompactEnabled": true,
  "autoCompactThreshold": 80,
  "criticalThreshold": 90
}

Cost calculation

The budget monitor now estimates input + output tokens separately and uses the model's real input/output prices. Example: Edit with a 200-char new_string counts as ~54 output tokens, charged at the model's output rate.

Effective Budget Multiplier

At 50%+ budget usage, the monitor shows how much CCO multiplies your effective budget — e.g. "1.6x more effective" if Read Cache + file digests saved enough redundant reads to make your 200K context behave like ~320K.