# Cost Awareness

> Activates when user asks about token usage or cost: "how much has this session cost", "what's my token usage", "am I close to my budget", "cost report", "how many tokens", "what have I spent", "token count", "check my usage". Also activates proactively when session shows very high token consumption: responses consistently exceeding 400 lines, more than 15 tool calls in one turn, or user mentions slow responses, context limits, or approaching compaction.

- Skill: `aarshsharma2803-stack/cost-awareness` (Agent Skill)
- Install (CLI): `npx skillmds@latest add aarshsharma2803-stack/cost-awareness`
- Raw SKILL.md: https://api.skillmd.com/api/skills/aarshsharma2803-stack/cost-awareness/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Docs & Writing
- Author: aarshsharma2803-stack (https://skillmd.com/u/aarshsharma2803-stack)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/aarshsharma2803-stack/cost-awareness

---


**Read `.optimusprime/cost-log.json`** (walk up from cwd to find `.optimusprime/`).

**Estimate current session tokens** from context window size (4 chars ≈ 1 token).

**Output:**
```
COST REPORT — Session: ~Xk in / ~Yk out | ~$Z.ZZ (Sonnet 4 pricing: $3/M in, $15/M out)
Previous sessions: N total | ~$W.WW cumulative
```

**Budget alert** based on `complexity_budget` in contract.json:
- `minimal` + usage > 20k tokens → warn
- `moderate` + usage > 60k tokens → warn
- `full` + usage > 150k tokens → warn

Alert: `COST ALERT: ~$X.XX used this session — over [minimal/moderate/full] budget`

**Append to `.optimusprime/cost-log.json`** (create if missing):
```json
{"session_id": "...", "estimated_input_tokens": N, "estimated_output_tokens": N, "estimated_cost_usd": N, "recorded_at": "ISO 8601"}
```