Q-Search is an automated, proof-gated research platform designed to investigate structural quantum algorithms for hard classical computational problems.
Rather than optimizing for small-scale demonstrations, toy circuits, or premature speedup claims on trivial instances, Q-Search systematically formulates high-upside quantum mechanisms, subjects them to automated classical attack suites, and permanently preserves negative results.
- Executive Summary
- Live Research Dashboards
- Repository Architecture & Codebase Layout
- Primary Research Tracks
- The Proof-Gated Operating System
- Getting Started & Quick Start
- Categorized CLI Command Reference
- Operating Contract & Model Allocation
The central question driving Q-Search is: Can genuine polynomial or super-polynomial quantum speedups be established for non-abelian hidden subgroup problems, linear code equivalence, or dihedral coset instances without falling to classical dequantization?
The latest audit rules out one specific target: a single bounded-norm,
inverse-polynomial-gap operator cannot completely label typical hidden-involution
multiplicity blocks. The same packing bound applies to any fixed number of
commuting operators. This is not an HSP or general circuit lower bound.
The next useful target is an adaptive coarse-label hierarchy or direct
source-aware transform, not another fitted global separator.
See the derivation, assumptions, and attempted refutations.
The follow-up signed-tensor access audit
also shows why the faithful diagram regime misses typical required K-types.
It leaves physical quotient representations and implicit copy registers open.
The encoded restriction workbench now checks
normalized two-QFT carrier extraction, noncommuting logical orbit averages,
and physical coset-register conventions. Its reference-orbit audit bounds
fixed-reference invariant identification, not binary class detection.
The reference-discard audit separately
bounds binary information after independent carrier/type discard. A joint
measurement on the remaining copy codes cannot recover the erased signal;
classical adaptive references still obey a weaker bound when fresh inputs
are twirled before memory interaction. A constructive coherent-reference
countercheck recovers the original input after carrier erasure by moving
its information first. This escapes that assumption, not the hard decoding
problem. Fully coherent access and one global reference twirl remain open.
The binary instrument evaluator
now measures actual outcome laws and disturbance. Its small-group alternating
carrier gains are reproduced by a latent-irrep model after a joint quantum
label front end; that front end has not been classically replaced. These
are calibration controls, not candidate algorithms.
Its instrument implementation audit
now checks clean GPE compute-copy-uncompute. Discarding the Fourier workspace
can preserve label probabilities while destroying the needed quantum state.
The clean label primitive is available conditionally on charged QFT/group-action
access; the growing-copy decision rule remains missing.
Exact fixed-point-free character arithmetic now scores supplied labels through
degree 4096 without tableau enumeration. Typed mutation search distinguishes
binary detection from hidden-element identification and retains one explicitly
incomplete growing-copy binary proposal. Known logarithmic-copy information
bounds are no longer listed as unresolved compiler requirements.
The first growing-copy route tested has a precise limitation: a fixed subset
palette exposes only one effective coset sample per membership-pattern cell.
This rules out amplifying raw copy counts with a fixed small palette. A
separate source-conditioned bound
also limits fixed preselected cell profiles when all classical source labels
are kept, by retaining small cells as fully charged quantum inputs.
It explicitly tests persistent conditional modes and missing-irrep measurement
extensions. Classical adaptation among a small predetermined whole-execution
catalogue is also bounded without normalizing selected successes. Succinct
exponential catalogues, coherent selectors and growing palettes remain outside
the useful bound; the derivation is review-pending, not an algorithm or a
novelty claim.
The coherent subset-phase query
now specifies an actual measurement family with polynomial primitive counts.
Independent source-conditioned kernels verify its finite readouts. The S4
signal beats one pair but loses to pair-plus-total; no scalable classifier
is established. A separate discarded-label bound does not apply to the
retained-source experiment, whose label/selector correlations matter.
Exact Walsh contractions now test larger copy counts, with disjoint-pair
baselines; S4's remaining gain uses a commuting V4-supported phase. A separate
review-pending bound limits fixed low-order selector marginals even with all
source labels, but leaves collective full-output classification open.
The terminal readout analysis now
implements explicit parity/threshold rules. Fixed parities reduce to random
two-subset tests; a source-uniform Fourier-envelope bound also limits the
corrected all-zero rule despite its finite success. Source-aware majority
remains outside that argument, not an established algorithmic advantage.
The source-selected parity test now
goes beyond that scope with exact character contractions through S6. Its
corrected negative-character rule loses to even one pair for every copy
count k>=2 at S6, certified by an exact finite prefix and a decreasing tail.
This is a scoped negative result, not a growing-degree no-go or a speedup.
The full-source Walsh analysis extends
the corrected-output bound: all source
irrep labels and complex central phases are now covered in the large
participating-copy regime, review pending. Exact joint laws expose source
information previously discarded and compare it with full pair likelihoods.
The intermediate copy window and multi-query measurements remain unresolved;
no scalable advantage, classical sampler or novelty claim is established.
The streamed S8 orbit probe now tests
the intermediate copy regime without a full group-pair matrix. Its ideal
sign-category/Walsh table loses to pair-count readouts at every tested point,
despite substantial information in the raw coset states. This eliminates
a specified finite readout, not the general intermediate regime.
The quantum-selector extension
now bounds the state before any final selector measurement, with classical
source labels retained and fixed-weight mask probabilities charged. It extends
to noncentral common group-algebra unitaries, provided one shared unknown
involution is averaged over its class AFTER tensor products. The
typical-mask refinement now gives a
review-pending obstruction for EVERY polynomial copy budget under this
architecture, including central-weight masks with charged overlap.
The environmental-fidelity extension
also obstructs masks with negligible mass below weight 3n at polynomial copy
counts, without any uniform-overlap penalty. Low-occupation masks, retained
physical data, source-adaptive operations and multiple queries remain
unresolved; all these derived bounds await independent and novelty review.
The mask-symmetry reduction narrows fixed-mask
information optimization to k+1 Hamming-weight probabilities while preserving
cross-weight coherence. It does not make the objective or final measurement
efficient, and its finite optimization controls are not algorithm candidates.
The vacuum-coherence bound further bounds
only the extra information from superposing the empty subset. At polynomial
copy budgets that gain is negligible under the stated architecture; mutual
coherences between nonempty low-weight masks remain unresolved. This is a
review-pending derivation, not a total-information bound or novelty claim.
The low-occupation support bound
bounds entire masks below a growing weight threshold after charging their
physical support dimension. For K=n^2, support at weights <=n/4 is obstructed
asymptotically. Separate sector bounds cannot be combined by discarding coherence.
The occupation-band localization
now charges the actual cross block for sufficiently separated low/high sectors.
At K=n^2 it makes nonnegligible mass between weights n/4 and 3n necessary
for a signal. The middle band and coherence with outside sectors remain
unresolved; this is not a sufficient condition or a working algorithm.
The source-resolved spin-block audit
preserves those coherences in an exact representation, but exposes its cost:
full irrep source labels are asymptotically distinct, leaving exponentially
large blocks under this symmetry alone. Coarsening sources can lose signal;
the reduction does not compile the remaining measurement.
The coefficient-mass channel bound
covers every fixed mask for sufficiently small group-algebra coefficient mass.
Polynomial-support queries and single low-dimensional irrep reflections
cannot escape through the intermediate band. Larger-mass compiled queries
and changed physical access remain unresolved.
The retained-data missing-harmonic detector
implements a costed measurement benchmark without a minimum-positive-gap
assumption. Its generic sqrt(n!) amplification is uncompetitive, not a
speedup. A scoped query-hybrid bound also rules out fixing it merely by
changing ancilla schedules while retaining the same missing-sign queries;
other data operations and implemented data readouts remain open.
The source-adaptive extension
also charges choosing subsets from full irrep labels, including the nonzero
information already in those labels. Other target sectors and noncentral
carrier operations remain outside these bounds.
Run python qsearch.py coset-missing-harmonic or
python qsearch.py run EXP-COSET-MISSING-HARMONIC-DETECTOR.
The coherent overlap echo now tests an
explicit retained-data sequence on nonmissing sectors, with a fixed X readout
and charged growing-copy resource schema. Exact S6 noncommuting controls lose
to the pair-plus-single baseline; growing-degree bias remains unproved.
Run python qsearch.py coset-overlap-echo or
python qsearch.py run EXP-COSET-COHERENT-OVERLAP-ECHO.
The exact spatial transfer evaluates
long one-round paths and certifies the fixed-S6 readout fails at every copy
count: even one pair measurement beats its global maximum. Growing degree
and multiple temporal rounds are not covered. Run
python qsearch.py coset-overlap-transfer; use --replay for exact saved-witness
and finite-prefix verification.
The physical DCP witness audit
retracts the assumption that reversing an arbitrary measurement prepares a
uniform subset-sum fiber. A full-space counterexample preserves valid witness
support but not uniformity. The repaired unamplified compute-copy-uncompute
reduction gives average witness success at least the square of the decoder's
label-averaged success, without estimating that success or assuming typical
fiber occupancy. This is a review-pending reduction, not a new decoder or a
hardness theorem. Uniform-fiber entanglement exclusions must not be imported
through it. Run python qsearch.py dcp-arbitrary-measurement-witness-reduction.
The classical value-access audit
replaces hidden full-table quadratic reconstruction with counted point queries,
including exact-residue controls on nonmaterializable domains. Sample-limited
exhaustive recovery is now distinguished from polynomial-time dequantization;
value-oracle attacks are not treated as attacks on DHSP phase states.
The same audit retracts the Fourier diagnostic's unsupported learner claims:
spectral concentration is not shift recovery, and superseded negatives are
preserved in a separate quarantine archive.
The proof-route audit now checks scoped dependency
contracts and pinned evidence. It catches unsupported assertions without
pretending to check mathematical truth or cover the entire repository.
Source-ranked S_14 scans remain numerical algebra diagnostics. Reports now retain generator matrices and separator coefficients for replay; caches are bound to the branch, numerical basis, and contraction source. Failed numerical searches remain inconclusive unless independently certified. No decoder or new quantum speedup has been established.
-
Toy Instance Illusions: Small circuits (
$n \le 3$ ) often show high simulated success rates that merely rediscover trivial parity relations (Bernstein-Vazirani) without scaling. - Classical Dequantization: Many heuristic quantum observables can be simulated efficiently by classical Information-Set Decoding (ISD), Weisfeiler-Leman (WL) graph refinements, sparse Fourier sampling, or lattice BDD heuristics.
- Erased Dead Ends: Unrecorded negative experiments lead subsequent research into cyclical, redundant investigations.
Q-Search enforces a strict claim-gating policy:
- Executable Research Checks: Modules contain derivations, finite diagnostics, and assumptions. Passing Python tests is not a machine-checked mathematical proof.
- Classical Baselines and Access Audits: Implemented attacks and model checks can falsify proposals; surviving them is not a classical lower bound.
- Negative Results: Scoped obstructions and failed hypotheses are retained in
research/registry/negative_results.json; record counts are not independent discoveries. - Speedup Claims Blocked: The registry actively gates
speedup_claim_allowed = Falseuntil a candidate provably defeats all named classical baselines across asymptotic families.
The repository automatically publishes interactive research telemetry and database dashboards via GitHub Pages:
| Dashboard | Description | Live Page |
|---|---|---|
| Progress Overview | Real-time candidate pipeline, falsifier telemetry, and validation status | index.html |
| Methodology | Formal research principles, no-go mechanisms, and claim-gating philosophy | methodology.html |
| Frontier Map | Machine-readable topological map of active research frontiers and kill criteria | frontier.html |
| Negative Results | Searchable database of 914 retained no-go theorems and dequantization findings | negative-results.html |
| Proof Debt | Live ledger of 24 open proof obligations, 1,184 lemmas, and reduction edges | proof-debt.html |
| Repository Map | Interactive codebase architecture explorer and 792-module taxonomy | repomap.html |
Q-Search contains 792 scientific verification modules. The codebase organizes modules into domain-specific packages (core/ and theorems/):
quantum-algorithm-search/
├── core/ # 22 Core Operating System & Proof Engine modules
│ ├── research_registry.py # Canonical schema for candidates, experiments & results
│ ├── experiment_runner.py # Experiment execution dispatcher and run history
│ ├── proof_gate.py # Formal candidate proof obligation and verification engine
│ ├── dequantization_checks.py # Automated classical attack matrix scanner
│ └── mutation_engine.py # Automated hypothesis mutation generator
│
├── theorems/ # 792 Scientific Theorem Verification modules
│ ├── dcp_*.py # Dihedral Coset Problem (DHSP) & state-native sieves
│ ├── coset_*.py, cfi_*.py # Non-abelian coset observables & S_n representation theory
│ ├── self_dual_wreath_*.py # Self-dual wreath product representations & polar audits
│ ├── code_*.py, goppa_*, bch_* # Linear code equivalence & automorphism baselines
│ └── character_*, phase_* # Phase family naturalness & Fourier bridge baselines
│
├── research/ # Canonical JSON registries & empirical attack artifacts
│ ├── registry/ # candidates.json, experiments.json, negative_results.json, etc.
│ ├── classical_baselines/ # Dequantization and classical attack outputs
│ ├── progress_snapshot.json # Curated telemetry feed powering public web dashboards
│ └── frontier_map.json # Structured research frontier topology
│
├── site/ # Frontend dashboard assets (styles.css, progress.js)
├── tools/ # Maintenance utilities (build_progress_snapshot.py, etc.)
├── docs/ # Human-readable repository maps and specifications
├── tests/ # Unit tests, integration tests, and runner dispatch suites
│
├── qsearch.py # The ONLY Python script at root (unified CLI entry point)
├── README.md # Modernized project guide
├── requirements.txt # Dependencies
└── [6 HTML Dashboards] # index.html, methodology.html, frontier.html, etc.
Benefits of this Architecture:
- Uncluttered Root: Only
qsearch.pyand configuration files reside at root. - Zero Packaging Friction: All 792 workflows execute seamlessly via
python3 qsearch.py <command>. - Clean Separation of Concerns: Core platform orchestration (
core/) is cleanly separated from domain theorem proofs (theorems/).
1. Dihedral Hidden Subgroup Problem (DHSP) & Sieve (DHS-GOWERS-SIEVE)
-
Objective: Recover hidden dihedral reflections from independent coset-state samples over
$D_N$ . -
Key Mechanisms: Uniform state-native sum/difference measurements, recursive multi-stage decoders, and bad-register contamination witnesses (
$1/\log N$ arbitrary error rate). -
Core Results & No-Go Theorems:
-
Lucas Carry ANF Invariance: Proved that arbitrary dense invertible affine Boolean preprocessing
$\text{GL}(m, 2)$ preserves linear algebraic-normal-form degree for 2-adic carries. - High-Quotient Distribution No-Go: Proved that low-only carry selection leaves the high quotient distribution generic, preventing shortcut lattice attacks without full joint constraints.
-
Lucas Carry ANF Invariance: Proved that arbitrary dense invertible affine Boolean preprocessing
- Objective: Test linear code equivalence over finite fields using collective coset observables.
- Key Mechanisms: Commutant algebras of symmetric group representations, Jucys-Murphy elements, Racah recoupling coefficients, and wreath product Hecke algebras.
- Core Results & No-Go Theorems:
- PGM Polar Boundary: Established exact quantum capacity limits on low-register tensor observables.
- Master Walsh Flatness No-Go: Proved that hyperoctahedral adaptive Walsh operators suffer exponential signal cancellation on regular orbits.
- Multiplicity Twirl Falsifier: Sparse signed-sector projection matches an exact hyperoctahedral twirl and shows selected rank-seven multiplicity-three and nontrivial-beta blocks close by moved-point support at most five, refuting strict support-growth extrapolation while leaving uniformity and coherent access open.
- Source-Weighted High-Mass Scan: A matrix-free signed-YJM fiber trace reaches the highest-mass previously untested repeated
S_14branch (b=26, source mass0.008705). Support three generates only dimension 7, while a support-four subset has direct common-commutant nullity one with next singular value0.223. Audited source coverage rises to0.008736, still below one percent; exact all-rank closure, gap scaling, coherent access, and decoding remain open.
- Objective: Investigate algebraic and combinatorial invariants beyond strong Fourier sampling.
- Key Mechanisms: Cai-Fürer-Immerman (CFI) gadget pairs, higher-order Weisfeiler-Leman invariants, and graphlet tensor contractions.
graph TD
A[Curated Literature & Ontologies] --> B[Hypothesis Formulation]
B --> C[Theorem Module Implementation]
C --> D[Classical Attack Matrix / Dequantization]
D -->|Classical Collision Found| E[Permanent Negative Result Record]
D -->|Survives Classical Attack| F[Proof Gate & Lemma Obligations]
F -->|Proof Debts Open| G[Active Candidate / Frontier]
F -->|Proof Complete & Asymptotic Separation| H[Speedup Claim Gate Passed]
E --> I[914 Retained No-Go Theorems]
- Python 3.11+ (Tested on Python 3.13)
- Node.js (for frontend static checking)
git clone https://github.com/Jaspersands/qsearch.git
cd qsearch
pip install -r requirements.txtRun a complete registry audit, check proof obligations, and validate data snapshots:
# 1. Run full registry and dequantization attack audit
python3 qsearch.py audit
# 2. Run dequantization scanner across all candidates
python3 qsearch.py dequantize
# 3. Validate entire research registry integrity (zero issues expected)
python3 qsearch.py validate
# 4. Rebuild the public dashboard progress snapshot
python3 tools/build_progress_snapshot.py# Run candidate-specific unit tests
python3 -m pytest tests/test_dcp_carry_affine_degree_invariance.py
# Run dispatch verification across experiment runners
python3 -m pytest tests/test_experiment_runner.pyAll 792 research workflows are accessible via python3 qsearch.py <subcommand>.
python3 qsearch.py audit # Full literature, hypothesis, and registry audit
python3 qsearch.py hypothesize # Generate proof-gated hypotheses from ontology
python3 qsearch.py dequantize # Run classical attack matrix and dequantization scan
python3 qsearch.py validate # Validate registry consistency and claim gates
python3 qsearch.py literature # Extract mechanisms from seed literature
python3 qsearch.py frontier # Rebuild and inspect research frontier topologyDihedral Coset Problem (DCP) & Phase Sieve Commands (Click to expand)
python3 qsearch.py dcp-samples --n-values 8,10,12 --sample-count 4096
python3 qsearch.py dcp-decode --n-values 8,10,12 --samples-per-stage 4096
python3 qsearch.py dcp-recurrence --n-values 8,12,16,20,24 --trials-per-point 12
python3 qsearch.py dcp-schedules --n-values 20,24,28,32 --budget-multiplier 2.0
python3 qsearch.py dcp-uniform-schedules --train-n-values 20,24,28 --unseen-n-values 32,36,40
python3 qsearch.py dcp-bad-registers --n-values 12,16,20,24
python3 qsearch.py dcp-contamination --n-values 8,10,12,14,16 --register-fractions 0.25,0.5,1.0
python3 qsearch.py dcp-witness-search --n-values 12,16,20,24 --maximum-weight 4
python3 qsearch.py dcp-clifford-witnesses --n-values 8,10,12,14,16
python3 qsearch.py dcp-clifford-contamination --n-values 6,8,10,12
python3 qsearch.py dcp-hadamard-scaling --n-values 6,8,10,12 --register-ratios 0.5,1.0,1.5,2.0
python3 qsearch.py dcp-random-decoder --n-values 8,10,12,14,16
python3 qsearch.py dcp-decoder-frontier
python3 qsearch.py dcp-multiscale-aliasing
python3 qsearch.py dcp-carry-high-part
python3 qsearch.py dcp-carry-affine-degree-invariance
python3 qsearch.py dcp-boolean-coset-separation
python3 qsearch.py dcp-marker-list-decoder
python3 qsearch.py dcp-marker-deviations
python3 qsearch.py dcp-marker-all-targets
python3 qsearch.py dcp-marker-vulnerable-coordinates
python3 qsearch.py dcp-marker-chart-union
python3 qsearch.py dcp-marker-target-beam
python3 qsearch.py dcp-pgm-gram-block-encoding
python3 qsearch.py dcp-pgm-qsvt-degree-obstruction
python3 qsearch.py dcp-coherent-fiber-erasure-boundary
python3 qsearch.py dcp-global-erasure-inversion-reduction
python3 qsearch.py dcp-approximate-erasure-coherence-reduction
python3 qsearch.py dcp-erasure-perturbation-reductionNon-Abelian Coset States & Symmetric Group Commands (Click to expand)
python3 qsearch.py coset-state
python3 qsearch.py coset-collective-search
python3 qsearch.py coset-pgm-capacity
python3 qsearch.py coset-holevo
python3 qsearch.py coset-stable-fourth-moment
python3 qsearch.py coset-stable-racah-spectrum
python3 qsearch.py coset-hidden-involution-binary-identification-self-reduction
python3 qsearch.py coset-hidden-involution-boundary-gauge-projection-commutation
python3 qsearch.py coset-hidden-involution-centralizer-fourier-transversal
python3 qsearch.py coset-hidden-involution-commutant-basis-closure
python3 qsearch.py coset-hidden-involution-degree-profile-subduction-uniqueness
python3 qsearch.py coset-hidden-involution-gelfand-tsetlin-chain-orthogonality
python3 qsearch.py coset-hidden-involution-hecke-generator-exchange-reduction
python3 qsearch.py coset-hidden-involution-highest-weight-multiplicity-separation
python3 qsearch.py coset-hidden-involution-pair-matching-charge-hierarchy
python3 qsearch.py coset-hidden-involution-racah-tensor-inversion-stability
python3 qsearch.py coset-hidden-involution-multiplicity-twirl-projection
python3 qsearch.py coset-hidden-involution-multiplicity-fiber-trace
python3 qsearch.py coset-hidden-involution-high-mass-support-scan
python3 qsearch.py coset-hidden-involution-source-weighted-support-portfolio
python3 qsearch.py coset-hidden-involution-spectral-label-budget
python3 qsearch.py coset-hidden-involution-signed-tensor-access
python3 qsearch.py coset-hidden-involution-encoded-restriction
python3 qsearch.py coset-hidden-involution-reference-twirl-information
python3 qsearch.py coset-binary-carrier-instruments
python3 qsearch.py proof-routes
python3 qsearch.py run EXP-COSET-HIDDEN-INVOLUTION-SPECTRAL-LABEL-BUDGET
python3 qsearch.py coset-hidden-involution-natural-support-six-mass-audit
python3 qsearch.py cfi-code-reduction
python3 qsearch.py cfi-structural-decoderSelf-Dual Wreath Product Representation Commands (Click to expand)
python3 qsearch.py self-dual-wreath-spectrum
python3 qsearch.py self-dual-wreath-hecke-audit
python3 qsearch.py self-dual-wreath-pgm-polar-audit
python3 qsearch.py self-dual-wreath-adaptive-walsh-support-concentration-reduction
python3 qsearch.py self-dual-wreath-mrs-identification-escape-theorem
python3 qsearch.py self-dual-wreath-carrier-subspace-invariance
python3 qsearch.py self-dual-wreath-gpe-fusion-tree-cs-boundary
python3 qsearch.py self-dual-wreath-hyperoctahedral-subduction-rigidity
python3 qsearch.py self-dual-wreath-regular-master-walsh-flatness-no-go
python3 qsearch.py self-dual-wreath-orientation-fixed-space-recoupling
python3 qsearch.py self-dual-wreath-racah-decoupling-gauge-uniqueness
python3 qsearch.py self-dual-wreath-source-adaptive-walsh-collision-reduction
python3 qsearch.py self-dual-wreath-trace-biased-adaptive-walsh-no-go
python3 qsearch.py self-dual-wreath-branch-character-cyclic-polar-compiler
python3 qsearch.py self-dual-wreath-branch-character-cyclic-quadrant-overlap
python3 qsearch.py self-dual-wreath-branch-character-equivariant-multiplier-normal-form
python3 qsearch.py self-dual-wreath-branch-character-gpe-dilation-separation
python3 qsearch.py self-dual-wreath-branch-character-label-coherent-power-map-boundary
python3 qsearch.py self-dual-wreath-branch-character-naimark-autocorrelation-fourier-boundary
python3 qsearch.py self-dual-wreath-branch-character-natural-frobenius-word-map
python3 qsearch.py self-dual-wreath-branch-character-polar-naimark-completion
python3 qsearch.py self-dual-wreath-branch-character-power-map-fourier-access-boundary
python3 qsearch.py self-dual-wreath-branch-character-raw-concentration-central-fourier-bridge
python3 qsearch.py self-dual-wreath-branch-character-raw-polar-matched-filter-boundary
python3 qsearch.py self-dual-wreath-branch-character-sector-resolved-whitening-no-go
python3 qsearch.py self-dual-wreath-branch-character-whole-sum-path-erasure-boundary
python3 qsearch.py self-dual-wreath-joint-character-analysis-map-normalization
python3 qsearch.py self-dual-wreath-joint-character-natural-sector-mass
python3 qsearch.py self-dual-wreath-joint-character-purification-access-boundary
python3 qsearch.py self-dual-wreath-orientation-kernel-character-tensor-boundary
python3 qsearch.py self-dual-wreath-orientation-kernel-hash-normalization-no-go
python3 qsearch.py self-dual-wreath-schur-branch-merger-polar-equivalence
python3 qsearch.py self-dual-wreath-schur-dilated-multiplicity-access
python3 qsearch.py self-dual-wreath-split-sector-branch-regularity
python3 qsearch.py self-dual-wreath-trace-biased-coefficient-rank-no-go
python3 qsearch.py self-dual-wreath-schur-companion-transform-scope-boundary
python3 qsearch.py self-dual-wreath-addressed-cross-map-pair-polar-gram-boundary
python3 qsearch.py self-dual-wreath-addressed-cross-map-linear-assembly-normalization-boundary
python3 qsearch.py self-dual-wreath-natural-q-scale-spectral-window-no-go
python3 qsearch.py self-dual-wreath-final-root-metric-access-width-no-go
python3 qsearch.py self-dual-wreath-final-root-scalar-mixer-no-go
python3 qsearch.py self-dual-wreath-final-root-byproduct-covariance-no-go
python3 qsearch.py self-dual-wreath-final-root-physical-preparation-extension-scope-boundary
python3 qsearch.py self-dual-wreath-final-root-program-contraction-normalization-no-go
python3 qsearch.py self-dual-wreath-final-root-purification-naimark-program-boundary
python3 qsearch.py self-dual-wreath-final-root-state-preparation-oracle-query-boundary
python3 qsearch.py self-dual-wreath-final-root-addressed-weyl-assembly-boundary
python3 qsearch.py self-dual-wreath-recursive-polar-normalization-conservation-boundary
python3 qsearch.py self-dual-wreath-affine-gpe-nodelocal-naimark-access-boundary
python3 qsearch.py self-dual-wreath-positive-naimark-access-equivalence-boundary
python3 qsearch.py self-dual-wreath-hierarchical-endpoint-schur-algebra-boundary
python3 qsearch.py self-dual-wreath-affine-flag-aggregate-schur-query-boundary
python3 qsearch.py self-dual-wreath-affine-node-frame-response-boundary
python3 qsearch.py self-dual-wreath-scale-free-endpoint-graph-transfer-boundary
python3 qsearch.py self-dual-wreath-cayley-endpoint-gauge-compiler
python3 qsearch.py self-dual-wreath-affine-star-cayley-compiler
python3 qsearch.py self-dual-wreath-pair-carrier-label-contextuality
python3 qsearch.py self-dual-wreath-occupied-carrier-octahedral-boundary
python3 qsearch.py self-dual-wreath-plancherel-carrier-contextuality
python3 qsearch.py self-dual-wreath-plancherel-carrier-nonidentity-tail
python3 qsearch.py self-dual-wreath-plancherel-carrier-near-derangement-reduction
python3 qsearch.py self-dual-wreath-plancherel-carrier-asymptotic-closure
python3 qsearch.py self-dual-wreath-plancherel-carrier-racah-access-boundary
python3 qsearch.py self-dual-wreath-carrier-noncentral-readout-boundary
python3 qsearch.py self-dual-wreath-carrier-conditioned-pgm-boundary
python3 qsearch.py self-dual-wreath-carrier-holevo-budget-theorem
python3 qsearch.py self-dual-wreath-carrier-branch-pgm-success-certificate
python3 qsearch.py self-dual-wreath-disjoint-pair-branch-pgm-compiler-boundary
python3 qsearch.py self-dual-wreath-disjoint-pair-covariance-polar-reduction
python3 qsearch.py self-dual-wreath-dimensionless-pgm-truncation-bridge
python3 qsearch.py self-dual-wreath-local-block-metric-normalization-no-goLinear Code Equivalence & Classical Attack Commands (Click to expand)
python3 qsearch.py code-equivalence
python3 qsearch.py goppa-codes
python3 qsearch.py reed-muller-codes
python3 qsearch.py cyclic-codes
python3 qsearch.py bch-codes
python3 qsearch.py quasi-cyclic-codes
python3 qsearch.py code-structural-invariants
python3 qsearch.py code-tuple-profile-baseline
python3 qsearch.py code-schur-filtration
python3 qsearch.py support-splitting-baseline
python3 qsearch.py information-set-decoding-baselineQ-Search strictly separates cognitive research roles to maximize scientific output and precision:
- High-Reasoning Models (Codex):
- Reserved for theorem derivations, mathematical mechanism formulation, representation-theoretic proofs, asymptotic complexity reductions, and decisive experiment design.
- Mechanical & Agentic Execution (Gemini / Antigravity):
- Dedicated to CLI subparser generation, registry upserts, unit test creation, snapshot rebuilding, website maintenance, and GitHub Actions CI synchronization.
No finite experiment or empirical success rate (