Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →brycewang-stanford/cav-supplementary
Use when deciding what belongs in a CAV (Computer Aided Verification) paper body versus its optional appendix and artifact, covering the LNCS page limits, the rule that reviewers are not obliged to read the appendix so decision-critical content stays in the body, where full proofs and benchmark tables live, and double-anonymous supplementary material for the anonymized categories.
Passed 20 of 21 test cases with the skill, and 15 of 21 without it.
and 3 more in the full report
5 fixed16 passed either way1 still failing0 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 5 test cases the skill fixed. Same model, same prompt, with and without the skill.
We are writing a CAV Industrial paper. We currently have 14 pages in our main body excluding references. Is our main body compliant with the CAV page limit for Industrial papers?
| Checked | Without skill | With skill |
|---|---|---|
| Identifies that CAV Industrial papers have a 10-page main body limit excluding references and appendices, making 14 pages non-compliant. | Missed | Passed |
The instructions your model reads when the skill loads.
Use this when assembling CAV supplementary material. The governing rule is strict and specific to CAV's format: a clearly marked appendix is allowed, but reviewers are not obliged to read it. So the paper must be judgeable from the body alone — anything a reviewer must see to accept lives inside the page limit, and the appendix and artifact only support it.
| Content | Body (within the page limit) | Appendix / artifact |
|---|---|---|
| The problem, the technique, the main theorem | Yes | — |
| Proof idea / sketch of the main result | Yes | Full formal proof |
| Supporting lemmas | Statement + role | Full proofs |
| Headline benchmark comparison + resource limits | Yes | Full per-instance tables |
| Soundness/differential-check result | The result + method | The raw logs and witnesses |
| Benchmark set + selection | Summary + revision | The full instance list, fetch scripts |
| Tool build/run instructions | A pointer | README + scripts (artifact) |
| Proof certificates / witnesses | A pointer | The certificates + a checker (artifact) |
If a reviewer would need to open the appendix or artifact to know whether the theorem holds or the comparison is fair, the paper is mis-partitioned — move that content into the body.
CAV's LNCS limits (Regular 18 pages; Short Tool / Application / Industrial 10 pages) exclude references and appendices, but the appendix is optional-to-read. Consequences:
[No identity] no authors, affiliations, acknowledgements, grants, cluster paths, or repo owners in
the appendix or the review-time artifact; anonymize the tool/solver name
test cases passed, vs 15 / 21 without
Published skills whose descriptions closely match this one, with their headline benchmark lift.
More skills from the same source and nearby in the registry, with their headline benchmark lift.