← ClaudeAtlas

cost-trackinglisted

Use when the user wants to capture AI tool costs, review spending trends, set cost budgets, or integrate cost data into health snapshots — guides quarterly cost capture, records data in a structured format, and updates MODEL_ROUTING.md with observed cost patterns
Habitat-Thinking/ai-literacy-superpowers · ★ 41 · Code & Development · score 66
Install: claude install-skill Habitat-Thinking/ai-literacy-superpowers
# Cost Tracking Capture, record, and track AI tool costs over time. Cost data feeds into health snapshots and informs model routing decisions. This skill does not access billing APIs directly — it guides the user through checking their provider dashboards and recording the data in a structured format that the plugin can read. ## Why Track Costs Without cost data: - MODEL_ROUTING.md routing rules are theoretical - The break-even calculation for self-hosting is impossible - Cost per PR/feature is unknown — no way to optimise - Budget conversations happen without evidence With cost data: - Routing rules are validated against actual spend - Self-hosting decisions are grounded in real numbers - Cost trends reveal whether AI adoption is efficient - Budget conversations have receipts ## Two actuals records: the quarterly snapshot and the per-PR record This skill owns **two** actuals formats under `observability/costs/`: - **The quarterly snapshot** (below) — a **provider-level aggregate** of spend and tokens across all work for a billing period, captured by `/cost-capture`. - **The per-PR actuals record** — a **single-task** structural footprint (which stages ran, review cycles, files) plus human-supplied token/cost figures when available, written by the integration-agent at merge time and read by the `cost-estimator` as a `kind: calibration` source. It lives under `observability/costs/per-pr/` and is defined in [`references/per-pr-actuals-format.md`](reference