技能创建器
skill-creator
创建、改进和评估自定义技能,把专业流程转化为可重复使用的 AI 能力。
topology-proof-validation【数据与研究】Use this quip-miner sub-skill for D-Wave topology files, solver ranges, topology hash/proof data, winning-solution validation, dumped BQMs, and topology troubleshooting.
独立 GitHub 快照 · 自动分类 · 以原作者文档为准
Use this sub-skill when the user needs to inspect QuIP topology files, validate topologies against D-Wave hardware, interpret solver h/J ranges, download/revalidate on-chain wins, dump BQMs, or diagnose topology/proof mismatches.
references/topology-and-proof-data.md for dwave_topologies, default/production topology targets, winning-solution archive commands, BQM dump shape, and energy/difficulty semantics.references/dwave-solver-ranges.md for h/J/extended-J/per-qubit coupling ranges and current QPoW headroom.references/troubleshooting.md for missing topology files, invalid subgraphs, topology hash mismatches, solver range overshoot, and live QPU safety.scripts/validate_topology_subgraph.py to compare a candidate topology JSON/GZip file against a reference topology file.python scripts/validate_topology_subgraph.py --reference dwave_topologies/topologies/advantage2_system1.json.gz --candidate path/to/topology.json.gz
python scripts/download_and_validate_wins.py --url wss://qpu-1.nodes.quip.network/rpc --max 50 --dump-bqm --out quip_wins
python scripts/dump_solver_ranges.py --stdout-only
The latter two bundled helpers may require network, chain access, or D-Wave credentials. Provide commands; do not run them without approval.
submit_proof stores a compact seed/proof, not the full BQM. --dump-bqm reconstructs Ising models from nonce + topology snapshot.h in {-1,0,+1} and binary J in {-1,+1}, inside standard D-Wave ranges.extended_j_range, per-qubit coupling guards, anneal schedules, reverse anneal, and normalization helpers are headroom/not currently used by QPoW.../mining-backends/SKILL.md.../telemetry-attempt-archive/SKILL.md.../maintainer-testing-release/SKILL.md.