Commit db4223e
proof: dangerous pattern closure across 4 submodules
Audit and fix of dangerous patterns (Obj.magic, postulate, Admitted)
across the nextgen-languages monorepo:
- julia-the-viper: eliminated Obj.magic with type-safe VS Code binding
- ephapax: added ctx_mark_used_lookup_at lemma, closed T_Var_Lin
consumption tracking admit, documented 4 remaining Admitted proofs
with dependency chains
- oblibeny: annotated 4 FFI/funext postulates as intentional axioms
with justification (cannot be proved — opaque C FFI boundary)
- phronesis: replaced 3 postulated operations with real implementations
(value equality, integer less-than, list membership)
Remaining dangerous patterns: 3 Admitted in ephapax/formal/Semantics.v
(deeply interconnected proof obligations requiring new helper lemmas),
4 intentional axioms in oblibeny (FFI boundary + function extensionality).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>1 parent 4bd450b commit db4223e
4 files changed
Lines changed: 4 additions & 4 deletions
Submodule julia-the-viper updated 1 file
- .github/workflows/codeql.yml+1-1
- .github/workflows/guix-nix-policy.yml+1-1
- .github/workflows/hypatia-scan.yml+2-2
- .github/workflows/mirror.yml+8-8
- .github/workflows/npm-bun-blocker.yml+1-1
- .github/workflows/quality.yml+3-3
- .github/workflows/rsr-antipattern.yml+1-1
- .github/workflows/scorecard.yml+2-2
- .github/workflows/secret-scanner.yml+4-4
- .github/workflows/security-policy.yml+1-1
- .github/workflows/semgrep.yml+1-1
- .github/workflows/ts-blocker.yml+1-1
- .github/workflows/wellknown-enforcement.yml+1-1
- .github/workflows/workflow-linter.yml+2-2
- src/abi/Crypto.idr+19-4
- src/abi/packages/hello/Interface.idr+3
0 commit comments