skills.rest
Jakeintech
JakeintechCOMMUNITY
Explore all skills by this author →
GitHub

forecast

Generates token budget requests from tracker work items and historical metered usage rates.

1|Updated Aug 17, 2026
Search Tags:#sprint planning#forecasting#jira#cost estimation#token budget#usage metering
One-click install
npx skills add https://github.com/Jakeintech/waybill --skill forecast-jakeintech
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: forecast
Source: https://github.com/Jakeintech/waybill/tree/main/skills/forecast
Command: npx skills add https://github.com/Jakeintech/waybill --skill forecast-jakeintech

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Engineers and teams struggle to justify AI token budget requests with credible numbers. This Skill produces a defensible token ask by multiplying upcoming committed work by the user's own historical token cost per unit of work, with an explicit planning buffer. ## Core Features & Use Cases - Historical rate computation: Derives tokens-per-point, hours-saved-per-point, and utilization percentages from metered waybill ledger data, never from manual estimates. - Upcoming work gathering: Pulls assigned sprint issues or epic children via Atlassian MCP, with a fallback to user-listed items flagged as self-estimated. - Honesty guardrails: Labels forecasts low confidence when fewer than 5 shipped stories have token data, keeps ranges as ranges, and refuses to invent rates without ledger evidence. - Use Case: Before a sprint planning meeting, ask for a token forecast and receive a committed-work table, a one-line ask with the 1.2 buffer stated, the statistical basis, projected hours saved, and risk framing. ## Quick Start Ask the assistant to forecast my token needs for next sprint based on my assigned Jira issues and waybill ledger history.

Frequently Asked Questions about forecast

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I forecast AI token needs for a sprint?▼

Run the forecast skill, which pulls your upcoming sprint issues via Atlassian MCP, computes your historical tokens-per-point rate from the waybill ledger, and multiplies committed points by that rate with a 1.2 planning buffer.

How is the tokens-per-point rate calculated?▼

The rate is the median over your most recent shipped stories that have both story points and metered token data, computed by the waybill engine via the query forecast command. Manual token fields act as overrides, not the source.

What happens if I have little historical token data?▼

If fewer than 5 shipped stories have token data, the forecast is labeled low confidence with an explanation. If the ledger has no token data at all, the skill proposes logging 2-3 sprints first and offers only a clearly labeled rough placeholder.

Does the forecast work without Jira or Atlassian access?▼

Yes, as a fallback you can list the work items manually. Items missing story points are marked as self-estimated in the output table rather than blocking the forecast.

Why does the forecast include a 1.2 buffer?▼

The buffer accounts for planning uncertainty and is always stated explicitly so you can adjust it. If utilization of your last grant was under roughly 70%, the skill recommends a smaller buffer or smaller ask to keep future requests credible.

Related Skills

Explore More Product & Management Skills→
tombeliebertombelieber
101

daily-cost

Monitor daily Claude Code spending with aggregate and per-session cost breakdowns.

Community
Intermediate
ronniepinnellronniepinnell

sprint

Review backlog items and prioritize work for upcoming sprints.

Community
Intermediate
PandaalaPandaala

ctl-architecture

Diagnose Edgion Gateway configuration sync mismatches across center, server, and client cache tiers.

Official
Intermediate
mferris77mferris77

cost

Calculate token-level costs for sessions and campaigns from Claude Code files.

Community
Advanced