agentleFS
odinfreeodinfree/storm-ecdsa-harness/skills/quantum-graded-divider-placement/SKILL.md

quantum-graded-divider-placement

Skill4 starsChanged 21 days agoMITView on GitHub

Price a reversible divider width cut by trading physically retained quotient tags against reversible reconstruction. Use when a parked passenger, reverse relay, stage join or final compiler pass determines whether the whole point addition gets narrower or cheaper.

The file

Files
SKILL.md

Trennwändli (quantum-graded-divider-placement)

Price the placed divider in both directions

Freeze the parent and its production compiler recipe. Mark the exact division spans and live input/output roles, including data parked outside the divider. Record the number of distinct physical IDs, the maximum numbered ID plus one required by the format, and the peak live roles. These answer different questions. Include bystanders and exterior transpositions when quoting whole-circuit width.

  1. Enumerate the real tags retained for reverse traversal. For each removed tag, specify how to regenerate it, its controls, donor, uncompute and phase obligations. Emit neighboring tag counts against the same parent. Require an explicit, in-range physical binding for every logical donor. A missing array lookup can silently turn into wire zero when written to a typed integer array.
  2. Join prelude, divider and follow-on arithmetic in each direction. Record logical-to-physical maps on both sides, signed frames, passenger transfers and the raw and post-pass byte hashes. A successful reverse relay is only a local width bound. Include admissions and releases at both ends in the wider control.
  3. Replace just the tested interiors in an otherwise stock full-width control. Compare parent and control on matched simulator inputs, including complete states at joins, phase, scratch and original failing IDs. Add a targeted large-shift Euclid input where the proposed window is tight. If a stock-successful input breaks, capture it and stop promoting that placement.
  4. Price the candidates after the actual stage compilers. Record raw records, compiled records, static CCX and executed Toffolis separately. Compute the mean executed difference across paired panels and its sample standard error. Repeat on the assembled target width, including exterior adapters and the stream's own native draw.

Raw record counts mislead: a pass can remove millions of raw records and almost nothing after the stage compilers. Carry a component delta as a component result until the assembled row has a measured width, executed cost and independent correctness result.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.