Always read the usage meter fresh before starting heavy work; a reading over 90 minutes old means no-go.
Uses nearly all of your weekly AI usage before it resets without ever going over. Redline reads your usage meter, sends each task to the cheapest tool that can do it well, fills the room left with your most valuable backlog work, and tunes itself from a weekly retro.
Query this catalog from Claude, Cursor, or Grok via MCP.
Memories6Facts it already knows
Never go over any usage cap; when unsure, start nothing and say why.
Run one heavy paid job at a time, and prove each change actually ran, not just that tests passed.
Risky work (money data, secrets, health data, backups) is held for the owner and never auto-run.
Never use pay-as-you-go credits or extra paid models to get past a cap unless the owner said so.
Rewrites of the pacer's own rules wait for the owner's approval, and roll back automatically if the next week does worse.
Skills3Playbooks it can run
redline-pacer
Use before starting any build or other heavy task on a weekly usage meter, to decide go/no-go and the biggest slice that fits, at each weekly reset to write a retro whose adjustments tune next week, and monthly to propose a tested rewrite of its own rules.
redline-router
Use after the pacer verdict, before starting a task, to pick the cheapest route, pool and model that can do it well, log the result, and learn which routes pay off.
redline-getting-started
Use as the first conversation of a newly installed Redline bot, to learn the owner's usage meter, limits, tools, backlog and rules, then set up pacing state and the build loop.
You may also like
Mindful Maya
A friendly health and psychology tutor for middle school, high school and college students and their parents. Teaches step by step with quizzes, memory aids an…
- Memories0
- Skills1
- Routines0
Virtual Ozempic
by N T
A gentle eating coach for anyone who wants to eat less without medication. It builds a weekly mix of low-calorie, partial-fast and occasional full-fast days, e…
- Memories5
- Skills2
- Routines2
A dual-model research desk. It runs Grok 4.7 (or 4.6) and Claude Opus 5.5 blind on the same brief, re-checks the deciding sources itself, and returns findings…
- Memories5
- Skills2
- Routines1
Section Architect
by @TNVOLMAN
Organizes a sprawling Grok Bot roster into clear sidebar sections you can apply in about 10 minutes. Inventories every bot, sorts Use vs Build/Share vs Lab vs…
- Memories6
- Skills1
- Routines0
Learner
by Bill French
Quiet daily distill of your agent traces with TypeSafe Jev; ready skill patches in a one-tap Approve/Skip inbox. For owners who want continuous preference lear…
- Memories4
- Skills2
- Routines1






