lean-arklib-proofs
Skill4 starsChanged 20 days agoMITView on GitHub
Formalize and verify mathematical or cryptographic claims in Lean 4 with ArkLib: exact-target builds, axiom audits, dependency checks, and the build evidence an evaluator expects.
In this repository34 other files it gives its agents.
- AGENTS.mdAGENTS.md
- accepted-proof-exact-retuneSkill
- beachhead-ratchetSkill
- circuit-edit-integration-safetySkill
- computer-algebra-oraclesSkill
- constraint-construction-searchSkill
- cross-model-review-gateSkill
- deepseek-pressure-testSkill
- ecdsafail-fastevalSkill
- ecdsafail-lowq-playbookSkill
- ecdsafail-oracle-loopSkill
- ecdsafail-pareto-ratchetSkill
- ecdsafail-scratch-dietSkill
- elicit-researchSkill
- exact-search-engineeringSkill
- fast-lean-proof-loopSkill
- hard-problem-contractSkill
- invariant-chartSkill
- jev-second-perspectiveSkill
- lean-finite-certificate-splittingSkill
- mission-commandSkill
- nasqret-playbookSkill
- pip-disciplineSkill
- portable-source-assertionsSkill
- proximity-prize-campaignSkill
- proximity-prize-frontier-archaeologySkill
- research-loopSkill
- route-compare-probe-disciplineSkill
- rust-archive-build-bindingSkill
- send-to-claude-fableSkill
- shot-autopsySkill
- squares-in-squares-campaignSkill
- verify-paired-native-measurementsSkill
- verify-research-transferSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
Read the governing project instructions, pinned Lean and ArkLib revisions, target file, build command, and acceptance rules.
- Very specific: line 22
Do not claim completion while sorry, admit, an unapproved axiom, or a placeholder theorem remains in the trusted dependency path.
- Very specific: line 23
Do not assume native_decide, code generation, or an external oracle is accepted.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Lean and ArkLib proofs
Convert an informal argument or computational finding into a theorem accepted by the target Lean 4 and ArkLib project. The exact repository target and its policy define acceptance.
Proof workflow
- Read the governing project instructions, pinned Lean and ArkLib revisions, target file, build command, and acceptance rules.
- State the theorem with the intended quantifiers, domains, side conditions, definitions, and normalization. Compare it directly with the external claim.
- Isolate computational evidence as a lemma hypothesis or re-prove it inside Lean. CAS, solver, or search output does not enter the trusted result merely because it guided the proof.
- Implement the smallest maintainable proof using existing project lemmas before adding custom infrastructure.
- Build the exact target in the authoritative worktree. Confirm that the build consumed the edited file rather than a shadow copy or stale artifact.
- Apply every relevant gate in references/acceptance-gates.md.
- Preserve theorem name, repository commit, toolchain lock, ArkLib revision, commands, and outputs.
Boundaries
- Do not claim completion while
sorry,admit, an unapproved axiom, or a placeholder theorem remains in the trusted dependency path. - Do not assume
native_decide, code generation, or an external oracle is accepted; follow the target repository's trust policy. - A successful editor check or unrelated package build does not replace the exact target build.
- A formal theorem can still formalize the wrong statement. Keep statement equivalence as a separate reviewed obligation.
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
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.

