[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7BQ5YFZT44SFCZDWZGS16S1":3,"skill-same:01M7BQ5YFZT44SFCZDWZGS16S1":45},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M7BQ5YFZT44SFCZDWZGS16S1","routing-subtour-elimination",null,"Subtour-elimination methods for TSP, VRP, pickup\u002Fdropoff routing, and routing MIPs with binary arc variables. Use when route-continuity constraints may permit disconnected cycles and the model needs MTZ constraints, flow-based connectivity constraints, DFJ subset cuts, or lazy\u002Fiterative subtour cuts.","linyouxue\u002Fskill-repair-benchmark","linyouxue","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F216151464?v=4","01M7BJEVCXKD74KH6FG9CBVSQB","林宥学","01M7BJEVE02NP217V93BKZFC0J",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fevaluation\u002Fdata\u002Fskillsbench87\u002Foriginals\u002Fbike-rebalance\u002Fskills\u002Frouting-subtour-elimination\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: routing-subtour-elimination\ndescription: Subtour-elimination methods for TSP, VRP, pickup\u002Fdropoff routing, and routing MIPs with binary arc variables. Use when route-continuity constraints may permit disconnected cycles and the model needs MTZ constraints, flow-based connectivity constraints, DFJ subset cuts, or lazy\u002Fiterative subtour cuts.\n---\n\n# Routing Subtour Elimination\n\nIn routing MIPs, degree and continuity constraints are not enough. A vehicle can have one depot-to-depot path and a separate closed cycle among stations. Add subtour-elimination constraints whenever binary arc variables decide routes.\n\nUse this base notation:\n\n```python\nSTART = \"depot_start\"\nEND = \"depot_end\"\nvehicles = range(K)\nstations = range(n)\nfrom_nodes = [START, *stations]\nto_nodes = [*stations, END]\narcs = [(i, j) for i in from_nodes for j in to_nodes if i != j and not (i == START and j == END)]\n\nx = {(v, i, j): model.addVar(vtype=\"B\", name=f\"x_{v}_{i}_{j}\") for v in vehicles for i, j in arcs}\n```\n\n## Required Base Route Constraints\n\nSubtour elimination assumes each selected station has matching inbound and outbound route arcs.\n\n```python\nfor v in vehicles:\n    model.addCons(quicksum(x[v, START, j] for j in stations) == 1)\n    model.addCons(quicksum(x[v, i, END] for i in stations) == 1)\n\n    for i in stations:\n        incoming = quicksum(x[v, j, i] for j in from_nodes if j != i)\n        outgoing = quicksum(x[v, i, j] for j in to_nodes if j != i)\n        model.addCons(incoming == outgoing)\n        model.addCons(outgoing \u003C= 1)\n```\n\nThe subtour methods below prevent station-only cycles that are disconnected from `START`.\n\n## 1. MTZ Order Constraints\n\nMTZ adds an order variable for each vehicle-station pair. If vehicle `v` travels from station `i` to station `j`, then `order[v, j]` must be greater than `order[v, i]`.\n\n```python\norder = {\n    (v, i): model.addVar(vtype=\"C\", lb=1, ub=max(1, n), name=f\"order_{v}_{i}\")\n    for v in vehicles\n    for i in stations\n}\n\nfor v in vehicles:\n    for i in stations:\n        for j in stations:\n            if i != j:\n                model.addCons(order[v, i] - order[v, j] + n * x[v, i, j] \u003C= n - 1)\n```\n\nPros:\n\n- Compact: `O(K n^2)` constraints and `O(K n)` extra variables.\n- Easy to implement in common Python optimization APIs.\n- Good default for small and medium benchmark instances.\n\nCons:\n\n- LP relaxation is weak compared with cutset or flow formulations.\n- Can be slow for larger VRPs.\n- Order variables are artificial; do not interpret them as service times unless you also model time.\n\nUse MTZ first when correctness and implementation speed matter more than best possible MIP strength.\n\n## 2. Single-Commodity Flow Connectivity\n\nAdd an artificial connectivity flow that starts at the depot and sends one unit to every visited station. This flow is not physical vehicle load.\n\n```python\nvisit = {\n    (v, i): quicksum(x[v, i, j] for j in to_nodes if j != i)\n    for v in vehicles\n    for i in stations\n}\n\nflow_arcs = [(i, j) for i in [START, *stations] for j in stations if i != j]\nf = {\n    (v, i, j): model.addVar(vtype=\"C\", lb=0, ub=n, name=f\"conn_flow_{v}_{i}_{j}\")\n    for v in vehicles\n    for i, j in flow_arcs\n}\n\nfor v in vehicles:\n    total_visits = quicksum(visit[v, i] for i in stations)\n    model.addCons(quicksum(f[v, START, j] for j in stations) == total_visits)\n\n    for i, j in flow_arcs:\n        model.addCons(f[v, i, j] \u003C= n * x[v, i, j])\n\n    for i in stations:\n        incoming_flow = quicksum(f[v, h, i] for h in [START, *stations] if h != i)\n        outgoing_flow = quicksum(f[v, i, j] for j in stations if j != i)\n        model.addCons(incoming_flow - outgoing_flow == visit[v, i])\n```\n\nPros:\n\n- Stronger connectivity logic than MTZ in many models.\n- Static constraints, no callback needed.\n- Works with optional station visits.\n\nCons:\n\n- Adds `O(K n^2)` continuous variables.\n- Do not reuse truck load as the connectivity flow when the vehicle can both pick up and drop off. Physical load can increase and decrease; connectivity flow should monotonically distribute artificial units.\n- More memory than MTZ.\n\nUse this when MTZ gives weak incumbents or slow progress and the instance is still modest in size.\n\n## 3. DFJ Subset Cuts\n\nFor every nonempty proper subset `S` of stations, selected station-to-station arcs inside `S` cannot form a closed cycle:\n\n```text\nsum_{i in S, j in S, i != j} x[v,i,j] \u003C= |S| - 1\n```\n\nFor very small `n`, static enumeration is possible:\n\n```python\nfrom itertools import combinations\n\nfor v in vehicles:\n    for r in range(2, n):\n        for S_tuple in combinations(stations, r):\n            S = set(S_tuple)\n            model.addCons(\n                quicksum(x[v, i, j] for i in S for j in S if i != j) \u003C= len(S) - 1\n            )\n```\n\nPros:\n\n- Strong, direct subtour elimination.\n- No artificial order or flow variables.\n\nCons:\n\n- Exponential number of constraints.\n- Static enumeration is only acceptable for small station counts.\n\nUse static DFJ only for tiny instances or as a debugging baseline.\n\n## 4. Lazy or Iterative Cut Separation\n\nThe strongest practical pattern is to solve with base route constraints, detect subtours in incumbents, add only the violated DFJ cuts, and continue.\n\nIn solvers with convenient lazy callbacks, add cuts during branch-and-bound. Some Python solver APIs require callback or constraint-handler plumbing for true lazy enforcement, so iterative cut separation is often simpler for portable benchmark code:\n\n```python\ndef selected_arcs(model, x, v, arcs):\n    return [(i, j) for i, j in arcs if model.getVal(x[v, i, j]) > 0.5]\n\n\ndef station_cycles_without_start(selected, stations):\n    succ = {i: j for i, j in selected}\n    cycles = []\n    seen = set()\n\n    for start in stations:\n        if start in seen or start not in succ:\n            continue\n        path = []\n        cur = start\n        pos = {}\n        while cur in succ and cur not in pos and cur not in seen:\n            pos[cur] = len(path)\n            path.append(cur)\n            cur = succ[cur]\n        seen.update(path)\n        if cur in pos:\n            cycle = path[pos[cur]:]\n            if START not in cycle and END not in cycle:\n                cycles.append(cycle)\n    return cycles\n\n\nwhile True:\n    model.optimize()\n    if model.getNSols() == 0:\n        raise RuntimeError(f\"no feasible solution; status={model.getStatus()}\")\n\n    cuts_added = 0\n    for v in vehicles:\n        selected = selected_arcs(model, x, v, arcs)\n        for cycle in station_cycles_without_start(selected, stations):\n            if len(cycle) >= 2:\n                S = set(cycle)\n                model.freeTransform()\n                model.addCons(\n                    quicksum(x[v, i, j] for i in S for j in S if i != j) \u003C= len(S) - 1\n                )\n                cuts_added += 1\n\n    if cuts_added == 0:\n        break\n```\n\nPros:\n\n- Adds only cuts that are needed.\n- Often stronger than MTZ.\n- Avoids exponential static SEC generation.\n\nCons:\n\n- Iterative resolve can be slower than a true callback.\n- Requires reliable subtour detection.\n- More moving parts than MTZ.\n\nUse this when static MTZ is too weak and the solver environment does not make lazy callbacks convenient.\n\n## Method Choice\n\n| Method | Best For | Avoid When |\n| --- | --- | --- |\n| MTZ | Quick, compact, small\u002Fmedium MIPs | Large hard VRPs where relaxation strength matters |\n| Single-commodity flow | Stronger static connectivity, optional visits | Memory is tight, or `n` is large |\n| Multi-commodity flow | Very strong small routing models | Most practical benchmark tasks; too many variables |\n| Static DFJ | Tiny instances, debugging | More than roughly 15-18 stations without careful filtering |\n| Lazy\u002Fiterative DFJ cuts | Strong routing models with many possible SECs | Solver API\u002Fcallback complexity is too risky |\n\nFor pickup\u002Fdropoff rebalancing, start with MTZ or artificial connectivity flow. Do not use physical truck load as the only subtour-elimination mechanism because pickup\u002Fdropoff load can increase and decrease and may not prove route connectivity.\n\n## Validation\n\nAfter solving, reconstruct each route by following selected arcs:\n\n```python\ndef extract_route(selected):\n    outgoing = dict(selected)\n    route = [START]\n    cur = START\n    seen = {START}\n    while cur != END:\n        if cur not in outgoing:\n            raise RuntimeError(f\"route disconnected at {cur!r}\")\n        cur = outgoing[cur]\n        if cur in seen and cur != END:\n            raise RuntimeError(f\"cycle detected at {cur!r}\")\n        route.append(cur)\n        seen.add(cur)\n    return route\n```\n\nFail fast if a selected solution has a disconnected cycle, repeated station, missing depot start, or missing depot end.\n","# Routing Subtour Elimination\n\nIn routing MIPs, degree and continuity constraints are not enough. A vehicle can have one depot-to-depot path and a separate closed cycle among stations. Add subtour-elimination constraints whenever binary arc variables decide routes.\n\nUse this base notation:\n\n```python\nSTART = \"depot_start\"\nEND = \"depot_end\"\nvehicles = range(K)\nstations = range(n)\nfrom_nodes = [START, *stations]\nto_nodes = [*stations, END]\narcs = [(i, j) for i in from_nodes for j in to_nodes if i != j and not (i == START and j == END)]\n\nx = {(v, i, j): model.addVar(vtype=\"B\", name=f\"x_{v}_{i}_{j}\") for v in vehicles for i, j in arcs}\n```\n\n## Required Base Route Constraints\n\nSubtour elimination assumes each selected station has matching inbound and outbound route arcs.\n\n```python\nfor v in vehicles:\n    model.addCons(quicksum(x[v, START, j] for j in stations) == 1)\n    model.addCons(quicksum(x[v, i, END] for i in stations) == 1)\n\n    for i in stations:\n        incoming = quicksum(x[v, j, i] for j in from_nodes if j != i)\n        outgoing = quicksum(x[v, i, j] for j in to_nodes if j != i)\n        model.addCons(incoming == outgoing)\n        model.addCons(outgoing \u003C= 1)\n```\n\nThe subtour methods below prevent station-only cycles that are disconnected from `START`.\n\n## 1. MTZ Order Constraints\n\nMTZ adds an order variable for each vehicle-station pair. If vehicle `v` travels from station `i` to station `j`, then `order[v, j]` must be greater than `order[v, i]`.\n\n```python\norder = {\n    (v, i): model.addVar(vtype=\"C\", lb=1, ub=max(1, n), name=f\"order_{v}_{i}\")\n    for v in vehicles\n    for i in stations\n}\n\nfor v in vehicles:\n    for i in stations:\n        for j in stations:\n            if i != j:\n                model.addCons(order[v, i] - order[v, j] + n * x[v, i, j] \u003C= n - 1)\n```\n\nPros:\n\n- Compact: `O(K n^2)` constraints and `O(K n)` extra variables.\n- Easy to implement in common Python optimization APIs.\n- Good default for small and medium benchmark instances.\n\nCons:\n\n- LP relaxation is weak compared with cutset or flow formulations.\n- Can be slow for larger VRPs.\n- Order variables are artificial; do not interpret them as service times unless you also model time.\n\nUse MTZ first when correctness and implementation speed matter more than best possible MIP strength.\n\n## 2. Single-Commodity Flow Connectivity\n\nAdd an artificial connectivity flow that starts at the depot and sends one unit to every visited station. This flow is not physical vehicle load.\n\n```python\nvisit = {\n    (v, i): quicksum(x[v, i, j] for j in to_nodes if j != i)\n    for v in vehicles\n    for i in stations\n}\n\nflow_arcs = [(i, j) for i in [START, *stations] for j in stations if i != j]\nf = {\n    (v, i, j): model.addVar(vtype=\"C\", lb=0, ub=n, name=f\"conn_flow_{v}_{i}_{j}\")\n    for v in vehicles\n    for i, j in flow_arcs\n}\n\nfor v in vehicles:\n    total_visits = quicksum(visit[v, i] for i in stations)\n    model.addCons(quicksum(f[v, START, j] for j in stations) == total_visits)\n\n    for i, j in flow_arcs:\n        model.addCons(f[v, i, j] \u003C= n * x[v, i, j])\n\n    for i in stations:\n        incoming_flow = quicksum(f[v, h, i] for h in [START, *stations] if h != i)\n        outgoing_flow = quicksum(f[v, i, j] for j in stations if j != i)\n        model.addCons(incoming_flow - outgoing_flow == visit[v, i])\n```\n\nPros:\n\n- Stronger connectivity logic than MTZ in many models.\n- Static constraints, no callback needed.\n- Works with optional station visits.\n\nCons:\n\n- Adds `O(K n^2)` continuous variables.\n- Do not reuse truck load as the connectivity flow when the vehicle can both pick up and drop off. Physical load can increase and decrease; connectivity flow should monotonically distribute artificial units.\n- More memory than MTZ.\n\nUse this when MTZ gives weak incumbents or slow progress and the instance is still modest in size.\n\n## 3. DFJ Subset Cuts\n\nFor every nonempty proper subset `S` of stations, selected station-to-station arcs inside `S` cannot form a closed cycle:\n\n```text\nsum_{i in S, j in S, i != j} x[v,i,j] \u003C= |S| - 1\n```\n\nFor very small `n`, static enumeration is possible:\n\n```python\nfrom itertools import combinations\n\nfor v in vehicles:\n    for r in range(2, n):\n        for S_tuple in combinations(stations, r):\n            S = set(S_tuple)\n            model.addCons(\n                quicksum(x[v, i, j] for i in S for j in S if i != j) \u003C= len(S) - 1\n            )\n```\n\nPros:\n\n- Strong, direct subtour elimination.\n- No artificial order or flow variables.\n\nCons:\n\n- Exponential number of constraints.\n- Static enumeration is only acceptable for small station counts.\n\nUse static DFJ only for tiny instances or as a debugging baseline.\n\n## 4. Lazy or Iterative Cut Separation\n\nThe strongest practical pattern is to solve with base route constraints, detect subtours in incumbents, add only the violated DFJ cuts, and continue.\n\nIn solvers with convenient lazy callbacks, add cuts during branch-and-bound. Some Python solver APIs require callback or constraint-handler plumbing for true lazy enforcement, so iterative cut separation is often simpler for portable benchmark code:\n\n```python\ndef selected_arcs(model, x, v, arcs):\n    return [(i, j) for i, j in arcs if model.getVal(x[v, i, j]) > 0.5]\n\n\ndef station_cycles_without_start(selected, stations):\n    succ = {i: j for i, j in selected}\n    cycles = []\n    seen = set()\n\n    for start in stations:\n        if start in seen or start not in succ:\n            continue\n        path = []\n        cur = start\n        pos = {}\n        while cur in succ and cur not in pos and cur not in seen:\n            pos[cur] = len(path)\n            path.append(cur)\n            cur = succ[cur]\n        seen.update(path)\n        if cur in pos:\n            cycle = path[pos[cur]:]\n            if START not in cycle and END not in cycle:\n                cycles.append(cycle)\n    return cycles\n\n\nwhile True:\n    model.optimize()\n    if model.getNSols() == 0:\n        raise RuntimeError(f\"no feasible solution; status={model.getStatus()}\")\n\n    cuts_added = 0\n    for v in vehicles:\n        selected = selected_arcs(model, x, v, arcs)\n        for cycle in station_cycles_without_start(selected, stations):\n            if len(cycle) >= 2:\n                S = set(cycle)\n                model.freeTransform()\n                model.addCons(\n                    quicksum(x[v, i, j] for i in S for j in S if i != j) \u003C= len(S) - 1\n                )\n                cuts_added += 1\n\n    if cuts_added == 0:\n        break\n```\n\nPros:\n\n- Adds only cuts that are needed.\n- Often stronger than MTZ.\n- Avoids exponential static SEC generation.\n\nCons:\n\n- Iterative resolve can be slower than a true callback.\n- Requires reliable subtour detection.\n- More moving parts than MTZ.\n\nUse this when static MTZ is too weak and the solver environment does not make lazy callbacks convenient.\n\n## Method Choice\n\n| Method | Best For | Avoid When |\n| --- | --- | --- |\n| MTZ | Quick, compact, small\u002Fmedium MIPs | Large hard VRPs where relaxation strength matters |\n| Single-commodity flow | Stronger static connectivity, optional visits | Memory is tight, or `n` is large |\n| Multi-commodity flow | Very strong small routing models | Most practical benchmark tasks; too many variables |\n| Static DFJ | Tiny instances, debugging | More than roughly 15-18 stations without careful filtering |\n| Lazy\u002Fiterative DFJ cuts | Strong routing models with many possible SECs | Solver API\u002Fcallback complexity is too risky |\n\nFor pickup\u002Fdropoff rebalancing, start with MTZ or artificial connectivity flow. Do not use physical truck load as the only subtour-elimination mechanism because pickup\u002Fdropoff load can increase and decrease and may not prove route connectivity.\n\n## Validation\n\nAfter solving, reconstruct each route by following selected arcs:\n\n```python\ndef extract_route(selected):\n    outgoing = dict(selected)\n    route = [START]\n    cur = START\n    seen = {START}\n    while cur != END:\n        if cur not in outgoing:\n            raise RuntimeError(f\"route disconnected at {cur!r}\")\n        cur = outgoing[cur]\n        if cur in seen and cur != END:\n            raise RuntimeError(f\"cycle detected at {cur!r}\")\n        route.append(cur)\n        seen.add(cur)\n    return route\n```\n\nFail fast if a selected solution has a disconnected cycle, repeated station, missing depot start, or missing depot end.\n",{"name":5,"description":7},"main","cdcb6d230f3b15c3c33f8a2bdd16439f43602cbe","evaluation\u002Fdata\u002Fskillsbench87\u002Foriginals\u002Fbike-rebalance\u002Fskills\u002Frouting-subtour-elimination",1,"https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark","https:\u002F\u002Fgithub.com\u002Flinyouxue","2026-11-14 20:28:35",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add linyouxue\u002Fskill-repair-benchmark --skill routing-subtour-elimination",{"tool":37,"command":38},"bunx","bunx skills add linyouxue\u002Fskill-repair-benchmark --skill routing-subtour-elimination",{"tool":40,"command":41},"pnpm","pnpm dlx skills add linyouxue\u002Fskill-repair-benchmark --skill routing-subtour-elimination","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fevaluation\u002Fdata\u002Fskillsbench87\u002Foriginals\u002Fbike-rebalance\u002Fskills\u002Frouting-subtour-elimination\u002FSKILL.md",true,[],[46,53,60,67,72,77],{"skillNo":47,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":50,"categories":51,"keywords":52,"publishedAt":19},"01M7E1HGVRY73NBJ022CJG13X6","video-frame-extraction","Extract frames from video files and save them as images using OpenCV","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_03\u002Finputs\u002Fskills\u002Fvideo-frame-extraction\u002FSKILL.md",[],[],{"skillNo":54,"name":55,"nameZh":6,"description":56,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":57,"categories":58,"keywords":59,"publishedAt":19},"01M7E1HAG5R7CJ9C8NYJS771MB","openai-vision","Analyze images and multi-frame sequences using OpenAI GPT vision models","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_03\u002Finputs\u002Fskills\u002Fopenai-vision\u002FSKILL.md",[],[],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":64,"categories":65,"keywords":66,"publishedAt":19},"01M7E1H7DADKB6QBWKC2KQ63SZ","image-ocr","Extract text content from images using Tesseract OCR via Python, including structured field extraction from receipts and forms","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_03\u002Finputs\u002Fskills\u002Fimage-ocr\u002FSKILL.md",[],[],{"skillNo":68,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"publishedAt":19},"01M7E1H0WBNT43TENGQY7AVZHD","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_02\u002Finputs\u002Fskills\u002Fvideo-frame-extraction\u002FSKILL.md",[],[],{"skillNo":73,"name":55,"nameZh":6,"description":56,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":76,"publishedAt":19},"01M7E1GTC7D4N4Q2EF03875J9E","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_02\u002Finputs\u002Fskills\u002Fopenai-vision\u002FSKILL.md",[],[],{"skillNo":78,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":19},"01M7E1GQ92CWFMGH53KSSE4WER","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_02\u002Finputs\u002Fskills\u002Fimage-ocr\u002FSKILL.md",[],[]]