Skip to content
DOCS

Eight problems, eight prompts.

Each row is a thing teams actually complain about, what Loupe does with it, and the exact rewrite that closed it.

EVERY TEAM

Rework loops

One task, nine prompts. Nobody sees it happen.

Loupe groups turns into tasks and flags loops over four.

−31% turns per task9 turns → 2 turns
BEFORE

fix the rate limiter

AFTER

Replace the fixed-window limiter in src/mw/rate_limit.go with a token bucket: 100 rps, burst 20. Tests must pass.

NEW JOINERS

Context blindness

31% of prompts name no file, branch or test.

A context score per prompt, plus a nudge in the editor.

2.3× first-try wins41 → 92 score
BEFORE

the login is broken

AFTER

Login fails on Safari 18 after src/auth/session.ts:88. Repro in e2e/auth.spec.ts.

SECURITY

Secrets in prompts

Keys and prod URLs pasted straight into chat.

Redaction on the laptop before anything is uploaded.

37 blocked last week0 secrets stored
BEFORE

connect to postgres://admin:hunter2@prod-db...

AFTER

connect to [redacted · database url]

MANAGERS

Ramp-up time

You cannot coach what you cannot see.

Week-one prompt quality vs team median, with two tips.

11 days → 6 days to first merge+34 score in week 2
BEFORE

how does this repo work

AFTER

Summarise the request path from cmd/api/main.go to the first handler. Cite files.

FINANCE

Silent spend

Agent bills grow, nobody can attribute them.

Cost per repo, team and merged PR.

$41k / yr recovered−58% cost per task
BEFORE

12-turn loop on infra · $3.40

AFTER

2-turn playbook run · $0.72

PLATFORM

Blind spots

A hook gets removed and telemetry quietly stops.

Fleet health, tamper alerts, cloud-agent enrolment.

96% → 100% coverage3 alerts open
BEFORE

4 laptops silent for 9 days

AFTER

MDM re-push, alert in Slack, 0 silent

TECH LEADS

Freestyle prompting

Everyone reinvents the same prompt badly.

Winning prompts become playbooks, pasted from the CLI.

68% of tasks start from a playbook9 playbooks live
BEFORE

14 versions of "write tests"

AFTER

1 playbook · 84% merge rate

COMPLIANCE

Audit questions

"Show me what AI touched this release."

Prompt → commit → PR trail, exportable.

Answer in 30 secondsSOC 2 evidence ready
BEFORE

a week of git archaeology

AFTER

one filtered export