{"id":"5a9ca84d-4638-450d-a8c9-c151e45cdb21","shortId":"HXNqpt","kind":"skill","title":"batch-plan-execute","tagline":"Use when the user wants AI to turn a requirement text, requirement document, and existing reviewed plan files into dependency-ordered implementation plans, a consolidated `checklist.md`, and, only after an explicit execution command, implement those plans with subagents in parall","description":"# Batch Plan Execute\n\nUse this skill when the task is to create, revise, retire, execute, or inspect module plans and the derived `checklist.md` based on a requirement source and the current `plans/` directory state.\n\nThis skill has two explicit modes:\n\n- `plan`: generate or revise implementation-ready plans and refresh `checklist.md`\n- `execute`: implement already-reviewed plans\n\nThis skill can:\n\n- accept requirement text from chat, a requirement file such as `requirements.md`, or an existing plan file or `plans/` directory\n- reconstruct the current planning state from the latest requirement source, existing plan files, and a hidden state file\n- detect mixed changes per module: new requirements, changed requirements, removed requirements, and review notes\n- build dependency-aware module or workstream plans that reflect serial and parallel work safely\n- generate a consolidated `checklist.md` from the latest requirement source plus authoritative review notes\n- execute approved plans with subagents only after an explicit execution command from the user\n\n## Inputs And Resolution\n\n### Primary Input Resolution\n\nChoose the primary input in this order:\n\n1. an explicit plan file path or `plans/` directory path mentioned by the user\n2. an explicit requirement file path mentioned by the user\n3. a file or document already attached in the task context\n4. the requirement text written directly in the chat\n\nUse these rules:\n\n- If a plan file path is provided, treat that plan lineage as the primary starting point and inspect the sibling `plans/` directory.\n- If a `plans/` directory path is provided, inspect only plan files and the state file inside that directory.\n- If a requirement file path is provided, read it directly and treat it as the latest requirement source.\n- If the input is raw chat text, treat the full user text as the latest requirement source.\n- If both a requirement source and a `plans/` directory are available, always use both.\n- Fail immediately if any referenced file or directory does not exist or is not readable.\n\n### Output Language Resolution\n\nUse these rules:\n\n- Match all user-visible output language to the dominant language of the latest requirement source.\n- If no requirement source exists, match the dominant language of the user's current-turn input.\n- If the user explicitly asks for translation or a specific output language, follow that instead.\n\n## Mode Routing\n\nDefault to `plan` mode.\n\nSwitch to `execute` mode only when the user explicitly issues an execution command in the current turn.\n\nExamples that DO resolve to `execute` mode:\n\n- `开始执行`\n- `按这个方案实现`\n- `去改代码`\n- `直接落地`\n- `implement now`\n- `apply the plan`\n\nExamples that DO NOT change mode from the default `plan` behavior:\n\n- `LGTM`\n- `OK`\n- `继续`\n- `按这个方向`\n- `review 过了`\n- any approval that confirms the plan but does not explicitly ask to start implementation\n\nUse these rules:\n\n- Review approval is not execution approval.\n- If the user is ambiguous, stay in `plan` mode and either refine the execution plan or ask for a clearer execution command.\n- Never infer execution intent from momentum, optimism, or lack of objections.\n\n## Shared Semantic Model\n\n### Requirement Preprocessing\n\nBefore extracting requirements, computing requirement-derived hashes, deriving checklist items, or interpreting review notes, preprocess the latest requirement source in this order:\n\n1. remove HTML comments outside fenced code blocks\n2. ignore Markdown structure inside fenced code blocks\n3. exclude document-frontmatter `---` delimiters from requirement segmentation\n4. detect standalone `---` lines in the remaining requirement body as requirement-item dividers only when they are outside fenced code blocks, outside removed comments, and surrounded by non-empty requirement content\n\nUse these rules:\n\n- Treat commented requirement content as nonexistent.\n- HTML comments inside fenced code blocks are not review notes.\n- Review comments are only meaningful inside `plans/*.md`; requirement sources use the same comment-boundary semantics for preprocessing only.\n- Empty divider segments, comment-only divider segments, and formatting-only divider changes must not create semantic requirement changes.\n\n### Requirement Item Identity And Segmentation\n\nUse these rules:\n\n- Treat each standalone `---`-delimited body segment as a distinct requirement item candidate before module mapping.\n- Treat each standalone `---`-delimited body segment as a new planning batch boundary, not merely as an extraction aid.\n- A new standalone `---`-delimited requirement segment must produce a new requirement item identity and a new plan lineage, even when its implementation area overlaps with an earlier module.\n- Do not merge a requirement item extracted from a later `---` segment back into an earlier plan lineage.\n- Requirement item state must preserve extraction-time identity, including `item_id`, `source_kind`, `source_order`, `source_excerpt_hash`, `body_hash_without_comments`, `mapped_module_slug`, and `planning_batch_id`.\n- Meta sections such as background, goals, scope, non-goals, assumptions, rollout, acceptance criteria, and non-functional requirements are global constraints unless they clearly describe implementation-bearing work.\n\n### Module Lineage And Ownership\n\nUse these rules:\n\n- Module boundaries should follow implementation ownership, shared code surface, migration boundaries, deployment boundaries, and owner-level deliverables rather than raw heading order.\n- Plan slugs should follow implementation ownership, not requirement heading text, when the two differ.\n- Merge requirement slices that land on the same implementation path, shared code surface, or owner-level deliverable.\n- Split only when the source clearly describes independent deliverables or strict prerequisite sequencing.\n- Do not create overlapping modules that both own the same shared change.\n- If a shared change is needed by multiple modules, assign one clear owner module lineage.\n- Resolve latest lineage artifacts by highest `rev`, otherwise latest base file.\n- Prefer slug matches first.\n- Allow rename matching only for safe one-to-one matches on `body_hash_without_heading`.\n- If rename matching is ambiguous, do not guess.\n\n### Dependency Layering And Parallelism\n\nAfter extracting candidate modules, derive an explicit dependency graph before classification or subagent spawning.\n\nUse these rules:\n\n- Classify each module as `foundation`, `feature`, or `follow-on`.\n- Record `depends_on`, `blocks`, `parallelizable_with`, and `shared_change_owner`.\n- Prefer parallel modules only when they do not require the same shared code ownership in the same phase.\n- Prefer serial sequencing when one module changes contracts, schemas, migrations, shared utilities, or infrastructure that another module consumes.\n- If the implementation is `A then B`, model that directly instead of forcing both into one peer list.\n- Do not infer fake parallelism. Sequence conservatively if safe ordering is unclear.\n\n## Artifact Model And File Contracts\n\n### Plan File Lineage\n\nWrite plan files into a `plans/` directory:\n\n- If the source is a requirement file, use a `plans/` folder next to that file.\n- If the source is a plan file or `plans/` directory, keep writing into that same `plans/` directory.\n- If the source is raw chat text, use `<cwd>/plans/`.\n\nCreate the `plans/` directory if it does not exist.\n\nBase plan files use this format:\n\n- `<module-slug>.md`\n\nRevision plan files use this format:\n\n- `<module-slug>.rev-<n>.md`\n\nUse these naming rules:\n\n- `new-plan` may create a base file if the module has no prior lineage.\n- `revise-plan` and `obsolete-plan` must always write a new `rev` file.\n- `no-op` must not write a new plan file.\n- Do not overwrite an existing file silently.\n- A new standalone `---`-delimited requirement segment must create a new base file and a new plan lineage.\n- Do not route a later `---`-delimited requirement segment into an earlier lineage, even if it touches the same implementation area or extends the same feature family.\n- Revisions apply only within the lineage created from that specific requirement segment after its base file exists.\n- Slugs derived from `---`-delimited requirement segments must still describe the implementation owner-level deliverable, not the raw segment order or divider label.\n\n### Checklist Contract\n\nThe derived checklist file must be:\n\n- `checklist.md`\n\nUse these rules:\n\n- If the latest requirement source is a readable file, write `checklist.md` next to that requirement file instead of inside `plans/`.\n- If no readable requirement file exists, fall back to writing `checklist.md` inside the active `plans/` directory.\n- Always overwrite the resolved `checklist.md` path in place when it can be refreshed safely.\n- `checklist.md` is a derived acceptance snapshot based on the latest requirement source and authoritative review notes.\n- It exists to help humans verify whether the requested outcomes and acceptance expectations have been satisfied.\n- It is not a plan lineage artifact, not an execution target, and not an implementation task list.\n- Review notes may refine or correct checklist content only when they change the authoritative requirement or acceptance intent.\n- If an earlier `checklist.md` exists, treat it only as an optional source of prior checkbox state for safely matched unchanged items, never as the source of truth for checklist content.\n\n### State File Contract\n\nThe hidden state file must be:\n\n- `.state.json`\n\nMaintain `plans/.state.json` as the planning baseline.\n\nThe state file must contain at least:\n\n- `requirement_source`\n- `requirement_hash`\n- `run_at`\n- `requirement_items`\n- `modules`\n\nEach module state object must contain at least:\n\n- `slug`\n- `title`\n- `status`\n- `source_excerpt_hash`\n- `body_hash_without_heading`\n- `requirement_item_ids`\n- `latest_plan_file`\n- `latest_rev`\n- `last_action`\n\nStatus values must be explicit:\n\n- `active`\n- `obsolete`\n\nUse these rules:\n\n- The state file describes planning lineage only.\n- Do not treat the state file as permission to execute.\n- If the state file is missing, bootstrap from the latest requirement source and the existing plan files instead of failing.\n- Bootstrap must reconstruct both requirement items and module mappings using the same shared semantic model as normal plan mode.\n- If the state file exists but is malformed and cannot be safely reconstructed from visible artifacts, fail.\n\n## Plan Mode\n\nUse [docs/plan.md](./docs/plan.md) as the detailed plan-writing and plan-refresh contract.\n\nPlan mode is responsible for:\n\n- extracting implementation-bearing modules from the latest requirement source under the shared semantic model\n- classifying each module into exactly one action: `new-plan`, `revise-plan`, `obsolete-plan`, or `no-op`\n- producing implementation-ready module plans\n- refreshing `checklist.md`\n- refreshing `plans/.state.json` after all affected modules are processed\n\nUse these rules:\n\n- Treat authoritative review notes from the latest plan lineage as input constraints for plan revisions.\n- Merge requirement changes and review notes into one revised module plan when both affect the same lineage.\n- Generate or refresh `checklist.md` during every successful `plan` run.\n- Keep detailed module extraction, checklist refresh, subagent planning, and plan assembly behavior in `docs/plan.md` rather than duplicating it here.\n\n## Execute Mode\n\nUse [docs/execute.md](./docs/execute.md) as the detailed execution contract.\n\nExecute mode is responsible for:\n\n- resolving the target plan file or `plans/` directory\n- selecting the latest reviewed artifact for each chosen module lineage\n- dispatching execution directly from approved plan boundaries and dependency metadata\n- validating the implemented change before reporting completion\n\nUse these rules:\n\n- Execute only after an explicit execution command in the current turn.\n- Execute the latest reviewed artifact for each selected module lineage.\n- Fail immediately if the latest target artifact still contains unresolved review comments.\n- Treat approved plans as the execution boundary source of truth. Do not re-scope modules, re-derive write ownership, or re-plan dependency layering during execute mode.\n- Keep target resolution, preflight, worker dispatch, integration, verification, and completion details in `docs/execute.md` rather than duplicating them here.\n\n## Failure Conditions\n\nStop and report the problem instead of generating weak output when:\n\n- the requirement file, plan file, or `plans/` directory is missing or unreadable\n- module extraction fails for a required requirement source\n- a plan would be mostly speculation with no meaningful codebase tie-in\n- multiple latest lineage candidates exist for one module and the ordering cannot be resolved safely\n- rename matching is ambiguous and cannot be reduced to a safe one-to-one mapping\n- shared changes or dependency ownership cannot be assigned to one clear module or workstream\n- the dependency graph contains cycles that cannot be justified as a single merged module\n- the state file exists but is malformed and cannot be safely reconstructed from visible artifacts\n\nDo not emit empty plan files, summary-only files, or a single aggregate plan when the task requires per-module outputs.","tags":["batch","plan","execute","skills","micooz","agent-skills"],"capabilities":["skill","source-micooz","skill-batch-plan-execute","topic-agent-skills"],"categories":["skills"],"synonyms":[],"warnings":[],"endpointUrl":"https://skills.sh/micooz/skills/batch-plan-execute","protocol":"skill","transport":"skills-sh","auth":{"type":"none","details":{"cli":"npx skills add micooz/skills","source_repo":"https://github.com/micooz/skills","install_from":"skills.sh"}},"qualityScore":"0.455","qualityRationale":"deterministic score 0.46 from registry signals: · indexed on github topic:agent-skills · 11 github stars · SKILL.md body (13,284 chars)","verified":false,"liveness":"unknown","lastLivenessCheck":null,"agentReviews":{"count":0,"score_avg":null,"cost_usd_avg":null,"success_rate":null,"latency_p50_ms":null,"narrative_summary":null,"summary_updated_at":null},"enrichmentModel":"deterministic:skill-github:v1","enrichmentVersion":1,"enrichedAt":"2026-05-18T19:08:11.778Z","embedding":null,"createdAt":"2026-04-22T13:03:33.881Z","updatedAt":"2026-05-18T19:08:11.778Z","lastSeenAt":"2026-05-18T19:08:11.778Z","tsv":"'/docs/execute.md':1722 '/docs/plan.md':1587 '/plans':1123 '1':212,557 '2':226,565 '3':236,573 '4':247,582 'accept':106,802,1347,1370,1408 'action':1499,1625 'activ':1326,1505 'affect':1651,1686 'aggreg':1985 'ai':10 'aid':714 'allow':937 'alreadi':100,241 'already-review':99 'alway':345,1175,1329 'ambigu':500,957,1916 'anoth':1035 'appli':453,1242 'approv':186,474,491,495,1755,1805 'area':737,1234 'artifact':925,1068,1381,1581,1745,1786,1798,1971 'ask':406,483,512 'assembl':1709 'assign':916,1936 'assumpt':800 'attach':242 'authorit':182,1356,1405,1659 'avail':344 'awar':160 'b':1044 'back':754,1320 'background':794 'base':69,931,1133,1158,1208,1255,1349 'baselin':1455 'batch':2,46,707,788 'batch-plan-execut':1 'bear':818,1607 'behavior':466,1710 'block':564,572,603,629,995 'bodi':590,686,701,779,949,1486 'bootstrap':1533,1547 'boundari':649,708,828,837,839,1757,1810 'build':157 'candid':693,967,1901 'cannot':1575,1909,1918,1934,1949,1965 'chang':145,150,460,667,673,906,910,1000,1026,1403,1675,1764,1930 'chat':110,255,322,1120 'checkbox':1424 'checklist':543,1281,1285,1398,1438,1703 'checklist.md':31,68,96,175,1289,1303,1323,1333,1343,1413,1646,1693 'choos':205 'chosen':1748 'classif':975 'classifi':982,1619 'clear':814,887,918,1939 'clearer':515 'code':563,571,602,628,834,875,1014 'codebas':1894 'command':38,195,435,517,1777 'comment':560,606,619,625,635,648,658,782,1803 'comment-boundari':647 'comment-on':657 'complet':1767,1843 'comput':537 'condit':1853 'confirm':476 'conserv':1062 'consolid':30,174 'constraint':811,1669 'consum':1037 'contain':1460,1477,1800,1946 'content':614,621,1399,1439 'context':246 'contract':1027,1072,1282,1442,1598,1727 'correct':1397 'creat':57,670,897,1124,1156,1205,1247 'criteria':803 'current':76,127,399,438,1780 'current-turn':398 'cycl':1947 'default':419,464 'delimit':578,685,700,718,1201,1220,1261 'deliver':844,881,890,1272 'depend':25,159,961,972,993,1759,1829,1932,1944 'dependency-awar':158 'dependency-ord':24 'deploy':838 'deriv':67,540,542,969,1259,1284,1346,1822 'describ':815,888,1266,1513 'detail':1590,1700,1725,1844 'detect':143,583 'differ':863 'direct':252,308,1047,1753 'directori':78,124,220,280,284,298,342,355,1082,1107,1114,1127,1328,1740,1872 'dispatch':1751,1839 'distinct':690 'divid':595,655,660,666,1279 'docs/execute.md':1721,1846 'docs/plan.md':1586,1712 'document':17,240,576 'document-frontmatt':575 'domin':378,392 'duplic':1715,1849 'earlier':741,757,1225,1412 'either':506 'emit':1974 'empti':612,654,1975 'even':733,1227 'everi':1695 'exact':1623 'exampl':440,456 'excerpt':777,1484 'exclud':574 'execut':4,37,48,60,97,185,194,425,434,445,494,509,516,520,1384,1526,1718,1726,1728,1752,1771,1776,1782,1809,1832 'exist':19,119,135,358,389,1132,1195,1257,1318,1360,1414,1541,1570,1902,1960 'expect':1371 'explicit':36,84,193,214,228,405,431,482,971,1504,1775 'extend':1236 'extract':535,713,749,766,966,1604,1702,1878 'extraction-tim':765 'fail':348,1546,1582,1792,1879 'failur':1852 'fake':1059 'fall':1319 'famili':1240 'featur':987,1239 'fenc':562,570,601,627 'file':22,113,121,137,142,216,230,238,262,291,295,302,353,932,1071,1074,1078,1089,1097,1104,1135,1142,1159,1180,1190,1196,1209,1256,1286,1301,1308,1317,1441,1446,1458,1495,1512,1522,1530,1543,1569,1737,1867,1869,1959,1977,1981 'first':936 'folder':1093 'follow':414,830,853,990 'follow-on':989 'forc':1050 'format':664,1138,1145 'formatting-on':663 'foundat':986 'frontmatt':577 'full':326 'function':807 'generat':87,172,1690,1861 'global':810 'goal':795,799 'graph':973,1945 'guess':960 'hash':541,778,780,950,1466,1485,1487 'head':848,858,952,1489 'help':1362 'hidden':140,1444 'highest':927 'html':559,624 'human':1363 'id':771,789,1492 'ident':676,727,768 'ignor':566 'immedi':349,1793 'implement':27,39,91,98,451,486,736,817,831,854,872,1040,1233,1268,1389,1606,1641,1763 'implementation-bear':816,1605 'implementation-readi':90,1640 'includ':769 'independ':889 'infer':519,1058 'infrastructur':1033 'input':199,203,208,319,401,1668 'insid':296,569,626,639,1311,1324 'inspect':62,276,288 'instead':416,1048,1309,1544,1859 'integr':1840 'intent':521,1409 'interpret':546 'issu':432 'item':544,594,675,692,726,748,761,770,1430,1470,1491,1552 'justifi':1951 'keep':1108,1699,1834 'kind':773 'label':1280 'lack':526 'land':868 'languag':364,375,379,393,413 'last':1498 'later':752,1219 'latest':132,178,314,331,382,551,923,930,1295,1352,1493,1496,1536,1611,1664,1743,1784,1796,1899 'layer':962,1830 'least':1462,1479 'level':843,880,1271 'lgtm':467 'line':585 'lineag':269,732,759,821,921,924,1075,1166,1214,1226,1246,1380,1515,1666,1689,1750,1791,1900 'list':1055,1391 'maintain':1450 'malform':1573,1963 'map':696,783,1555,1928 'markdown':567 'match':369,390,935,939,947,955,1428,1914 'may':1155,1394 'md':641,1139,1147 'meaning':638,1893 'mention':222,232 'mere':710 'merg':745,864,1673,1955 'meta':790 'metadata':1760 'migrat':836,1029 'miss':1532,1874 'mix':144 'mode':85,417,422,426,446,461,504,1565,1584,1600,1719,1729,1833 'model':531,1045,1069,1561,1618 'modul':63,147,161,695,742,784,820,827,899,915,920,968,984,1004,1025,1036,1162,1471,1473,1554,1608,1621,1643,1652,1682,1701,1749,1790,1819,1877,1905,1940,1956,1993 'momentum':523 'most':1889 'multipl':914,1898 'must':668,721,763,1174,1184,1204,1264,1287,1447,1459,1476,1502,1548 'name':1150 'need':912 'never':518,1431 'new':148,705,716,724,730,1153,1178,1188,1199,1207,1212,1627 'new-plan':1152,1626 'next':1094,1304 'no-op':1181,1636 'non':611,798,806 'non-empti':610 'non-funct':805 'non-goal':797 'nonexist':623 'normal':1563 'note':156,184,548,633,1358,1393,1661,1678 'object':528,1475 'obsolet':1172,1506,1633 'obsolete-plan':1171,1632 'ok':468 'one':917,944,946,1024,1053,1624,1680,1904,1925,1927,1938 'one-to-on':943,1924 'op':1183,1638 'optim':524 'option':1420 'order':26,211,556,775,849,1065,1277,1908 'otherwis':929 'outcom':1368 'output':363,374,412,1863,1994 'outsid':561,600,604 'overlap':738,898 'overwrit':1193,1330 'owner':842,879,919,1001,1270 'owner-level':841,878,1269 'ownership':823,832,855,1015,1824,1933 'paral':45 'parallel':169,964,1003,1060 'paralleliz':996 'path':217,221,231,263,285,303,873,1334 'peer':1054 'per':146,1992 'per-modul':1991 'permiss':1524 'phase':1019 'place':1336 'plan':3,21,28,41,47,64,77,86,93,102,120,123,128,136,164,187,215,219,261,268,279,283,290,341,421,455,465,478,503,510,640,706,731,758,787,850,1073,1077,1081,1092,1103,1106,1113,1126,1134,1141,1154,1169,1173,1189,1213,1312,1327,1379,1454,1494,1514,1542,1564,1583,1592,1596,1599,1628,1631,1634,1644,1665,1671,1683,1697,1706,1708,1736,1739,1756,1806,1828,1868,1871,1886,1976,1986 'plan-refresh':1595 'plan-writ':1591 'plans/.state.json':1451,1648 'plus':181 'point':274 'prefer':933,1002,1020 'preflight':1837 'preprocess':533,549,652 'prerequisit':893 'preserv':764 'primari':202,207,272 'prior':1165,1423 'problem':1858 'process':1654 'produc':722,1639 'provid':265,287,305 'rather':845,1713,1847 'raw':321,847,1119,1275 're':1817,1821,1827 're-der':1820 're-plan':1826 're-scop':1816 'read':306 'readabl':362,1300,1315 'readi':92,1642 'reconstruct':125,1549,1578,1968 'record':992 'reduc':1920 'referenc':352 'refin':507,1395 'reflect':166 'refresh':95,1341,1597,1645,1647,1692,1704 'remain':588 'remov':152,558,605 'renam':938,954,1913 'report':1766,1856 'request':1367 'requir':14,16,72,107,112,133,149,151,153,179,229,249,301,315,332,337,383,387,532,536,539,552,580,589,593,613,620,642,672,674,691,719,725,747,760,808,857,865,1010,1088,1202,1221,1251,1262,1296,1307,1316,1353,1406,1463,1465,1469,1490,1537,1551,1612,1674,1866,1882,1883,1990 'requirement-deriv':538 'requirement-item':592 'requirements.md':116 'resolut':201,204,365,1836 'resolv':443,922,1332,1733,1911 'respons':1602,1731 'retir':59 'rev':928,1146,1179,1497 'review':20,101,155,183,471,490,547,632,634,1357,1392,1660,1677,1744,1785,1802 'revis':58,89,1140,1168,1241,1630,1672,1681 'revise-plan':1167,1629 'rollout':801 'rout':418,1217 'rule':258,368,489,617,681,826,981,1151,1292,1509,1657,1770 'run':1467,1698 'safe':171,942,1064,1342,1427,1577,1912,1923,1967 'satisfi':1374 'schema':1028 'scope':796,1818 'section':791 'segment':581,656,661,678,687,702,720,753,1203,1222,1252,1263,1276 'select':1741,1789 'semant':530,650,671,1560,1617 'sequenc':894,1022,1061 'serial':167,1021 'share':529,833,874,905,909,999,1013,1030,1559,1616,1929 'sibl':278 'silent':1197 'singl':1954,1984 'skill':51,81,104 'skill-batch-plan-execute' 'slice':866 'slug':785,851,934,1258,1480 'snapshot':1348 'sourc':73,134,180,316,333,338,384,388,553,643,772,774,776,886,1085,1100,1117,1297,1354,1421,1434,1464,1483,1538,1613,1811,1884 'source-micooz' 'spawn':978 'specif':411,1250 'specul':1890 'split':882 'standalon':584,684,699,717,1200 'start':273,485 'state':79,129,141,294,762,1425,1440,1445,1457,1474,1511,1521,1529,1568,1958 'state.json':1449 'status':1482,1500 'stay':501 'still':1265,1799 'stop':1854 'strict':892 'structur':568 'subag':43,189,977,1705 'success':1696 'summari':1979 'summary-on':1978 'surfac':835,876 'surround':608 'switch':423 'target':1385,1735,1797,1835 'task':54,245,1390,1989 'text':15,108,250,323,328,859,1121 'tie':1896 'tie-in':1895 'time':767 'titl':1481 'topic-agent-skills' 'touch':1230 'translat':408 'treat':266,310,324,618,682,697,1415,1519,1658,1804 'truth':1436,1813 'turn':12,400,439,1781 'two':83,862 'unchang':1429 'unclear':1067 'unless':812 'unread':1876 'unresolv':1801 'use':5,49,256,346,366,487,615,644,679,824,979,1090,1122,1136,1143,1148,1290,1507,1556,1585,1655,1720,1768 'user':8,198,225,235,327,372,396,404,430,498 'user-vis':371 'util':1031 'valid':1761 'valu':1501 'verif':1841 'verifi':1364 'visibl':373,1580,1970 'want':9 'weak':1862 'whether':1365 'within':1244 'without':781,951,1488 'work':170,819 'worker':1838 'workstream':163,1942 'would':1887 'write':1076,1109,1176,1186,1302,1322,1593,1823 'written':251 '去改代码':449 '开始执行':447 '按这个方向':470 '按这个方案实现':448 '直接落地':450 '继续':469 '过了':472","prices":[{"id":"f443dd34-ab4f-49c2-aae5-677517d58fa3","listingId":"5a9ca84d-4638-450d-a8c9-c151e45cdb21","amountUsd":"0","unit":"free","nativeCurrency":null,"nativeAmount":null,"chain":null,"payTo":null,"paymentMethod":"skill-free","isPrimary":true,"details":{"org":"micooz","category":"skills","install_from":"skills.sh"},"createdAt":"2026-04-22T13:03:33.881Z"}],"sources":[{"listingId":"5a9ca84d-4638-450d-a8c9-c151e45cdb21","source":"github","sourceId":"micooz/skills/batch-plan-execute","sourceUrl":"https://github.com/micooz/skills/tree/main/skills/batch-plan-execute","isPrimary":false,"firstSeenAt":"2026-04-22T13:03:33.881Z","lastSeenAt":"2026-05-18T19:08:11.778Z"}],"details":{"listingId":"5a9ca84d-4638-450d-a8c9-c151e45cdb21","quickStartSnippet":null,"exampleRequest":null,"exampleResponse":null,"schema":null,"openapiUrl":null,"agentsTxtUrl":null,"citations":[],"useCases":[],"bestFor":[],"notFor":[],"kindDetails":{"org":"micooz","slug":"batch-plan-execute","github":{"repo":"micooz/skills","stars":11,"topics":["agent-skills"],"license":"mit","html_url":"https://github.com/micooz/skills","pushed_at":"2026-04-08T05:00:46Z","description":"Agent Skills I created and use most often here.","skill_md_sha":"4f88f71a43ba2f2848ea7c2826a6ad23e58d4c08","skill_md_path":"skills/batch-plan-execute/SKILL.md","default_branch":"main","skill_tree_url":"https://github.com/micooz/skills/tree/main/skills/batch-plan-execute"},"layout":"multi","source":"github","category":"skills","frontmatter":{"name":"batch-plan-execute","description":"Use when the user wants AI to turn a requirement text, requirement document, and existing reviewed plan files into dependency-ordered implementation plans, a consolidated `checklist.md`, and, only after an explicit execution command, implement those plans with subagents in parallel where safe and in sequence where required."},"skills_sh_url":"https://skills.sh/micooz/skills/batch-plan-execute"},"updatedAt":"2026-05-18T19:08:11.778Z"}}