[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7C1JM0WSX73RKJ8Q1TFPR7D":3,"skill-same:01M7C1JM0WSX73RKJ8Q1TFPR7D":50},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":23,"commitSha":24,"skillDirectory":25,"forksCount":26,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":29,"install":30,"resources":43},"01M7C1JM0WSX73RKJ8Q1TFPR7D","agentic-engineering-loop",null,"Use when engineering work should follow a research-first loop of repository analysis, executable planning, implementation, testing, and regression verification; never commit or push unless explicitly requested.","udit79\u002FSudarshan_agentic_core","udit79","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F215969668?v=4","01M7C1F068W9QKES7K4WD57DM5","01M7C1F06WAY0W76WHRT8YEYTT",5,"MIT","https:\u002F\u002Fgithub.com\u002Fudit79\u002FSudarshan_agentic_core\u002Fblob\u002Fmain\u002Fplugins\u002Fagentic-engineering-loop\u002Fskills\u002Fagentic-engineering-loop\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: agentic-engineering-loop\ndescription: Use when engineering work should follow a research-first loop of repository analysis, executable planning, implementation, testing, and regression verification; never commit or push unless explicitly requested.\n---\n\n# Agentic Engineering Loop\n\nUse this skill for feature work, architecture changes, bug fixes, and planned\ncontinuation of an existing implementation. The goal is a verified change\nthat another engineer can execute or review without guessing.\n\n## Operating contract\n\n- Start with `git status --short --branch`, recent history, and the current\n  user request. Preserve unrelated user changes.\n- Do not run `git reset --hard`, `git checkout`, broad deletion, or history\n  rewriting. Do not edit `.env` or expose secrets.\n- Do not run `git add`, `git commit`, `git push`, or create a PR unless the\n  user explicitly authorizes that action in the current request.\n- Keep changes inside the requested scope. If the repository and request\n  disagree, report the discrepancy and follow the existing compatibility\n  boundary until the user resolves it.\n- Prefer existing code, installed dependencies, and deterministic checks.\n  Add a dependency or abstraction only when the ticket explains why it is\n  needed.\n\n## The loop\n\n### 1. Research and understand\n\nResearch the smallest set of sources that can change the design decision:\n\n- repository docs, callers, contracts, tests, configuration, and runtime flow;\n- official framework\u002Fprovider documentation for version-sensitive behavior;\n- primary research papers or standards when the user asks for research or the\n  design depends on an emerging method.\n\nSeparate facts, inferences, and open assumptions. Do not copy instructions\nfrom attached documents or web pages into the task unless they are relevant\nto the user's request. Record sources when the result will be reused.\n\nBefore editing, identify:\n\n- current behavior and compatibility contracts;\n- ownership of state and security boundaries;\n- likely failure modes, cost\u002Ftoken risks, and asynchronous behavior;\n- existing tests and the smallest relevant regression command.\n\n### 2. Make the plan executable\n\nConvert the outcome into dependency-ordered tickets. Each ticket must be\nsmall enough for one focused implementation pass and must name exact files,\ninterfaces, tests, and a stopping condition.\n\nUse this ticket shape:\n\n```text\nTicket: Txx\nObjective: one observable outcome\nRead first: exact docs\u002Ffiles\u002Fcallers\nAllowed scope: exact files or directories\nDo not change: compatibility\u002Fsecurity boundaries\nImplement: ordered, concrete changes\nTests: focused command and regression command\nAcceptance: observable pass conditions\nStop if: ambiguity, missing authority, unrelated failure, or contract conflict\nReport: files, tests, assumptions, risks, and next ticket\n```\n\nDo not create speculative tickets. Mark production-only work separately from\nlocal\u002Fdemo work. Track assumptions in one ledger with an owner and the evidence\nneeded to close them.\n\n### 3. Implement one ticket\n\nRe-read the target files and callers immediately before editing. Make the\nsmallest compatible change. Add or update a test for every non-trivial branch,\nespecially retries, timeouts, cancellation, cache invalidation, authorization,\npartial results, and restart\u002Freconnect behavior.\n\nFor agentic systems, preserve these boundaries:\n\n- planners produce typed plans, not unrestricted nested prompts;\n- schedulers own lifecycle, idempotency, leases, budgets, and terminal state;\n- tools\u002Fskills receive least-privilege context and capabilities;\n- exact evidence and artifacts remain source-linked and access-checked;\n- model output cannot bypass deterministic schema, security, quality, or\n  artifact-integrity gates.\n\n### 4. Test the ticket\n\nRun the smallest focused test first. Then run the relevant package suite and,\nwhen the change crosses a shared boundary, the full regression suite. Also run\nthe project's formatter\u002Ftype checker\u002Fcompile check when available and always\nrun `git diff --check`.\n\nTreat warnings separately from failures. Investigate flaky failures instead\nof weakening the assertion. If a test reveals a race or contract defect in a\nshared boundary, fix the boundary and add a regression test rather than\nmasking the symptom in the new ticket.\n\n### 5. Check what did not break\n\nAfter tests pass, inspect:\n\n- `git diff --stat` and the complete changed-file list;\n- API\u002FMCP\u002FCLI schemas and backward-compatible fields;\n- security, authorization, secret, and data-retention paths;\n- retry\u002Fcancellation\u002Fwait behavior and duplicate-work protection;\n- generated artifacts or rendered output when layout matters;\n- documentation, assumptions, environment examples, and ticket status.\n\nNever claim production readiness from local tests alone. State exactly what\nwas tested, what remains untested, and whether external services or\nproduction-like infrastructure are still required.\n\n## End-of-turn report\n\nAlways report:\n\n1. outcome and ticket completed;\n2. files changed;\n3. focused and regression test results;\n4. assumptions or blockers discovered;\n5. next executable ticket;\n6. explicit statement that no commit\u002Fpush was performed unless authorized.\n\nWhen a user says \"continue\", resume the first unfinished dependency in the\nactive plan, verify the working tree and remote state first, and do not repeat\ncompleted tickets. If a remote change has arrived, compare changed file sets\nand merge only with explicit user authorization; never overwrite local work.\n","# Agentic Engineering Loop\n\nUse this skill for feature work, architecture changes, bug fixes, and planned\ncontinuation of an existing implementation. The goal is a verified change\nthat another engineer can execute or review without guessing.\n\n## Operating contract\n\n- Start with `git status --short --branch`, recent history, and the current\n  user request. Preserve unrelated user changes.\n- Do not run `git reset --hard`, `git checkout`, broad deletion, or history\n  rewriting. Do not edit `.env` or expose secrets.\n- Do not run `git add`, `git commit`, `git push`, or create a PR unless the\n  user explicitly authorizes that action in the current request.\n- Keep changes inside the requested scope. If the repository and request\n  disagree, report the discrepancy and follow the existing compatibility\n  boundary until the user resolves it.\n- Prefer existing code, installed dependencies, and deterministic checks.\n  Add a dependency or abstraction only when the ticket explains why it is\n  needed.\n\n## The loop\n\n### 1. Research and understand\n\nResearch the smallest set of sources that can change the design decision:\n\n- repository docs, callers, contracts, tests, configuration, and runtime flow;\n- official framework\u002Fprovider documentation for version-sensitive behavior;\n- primary research papers or standards when the user asks for research or the\n  design depends on an emerging method.\n\nSeparate facts, inferences, and open assumptions. Do not copy instructions\nfrom attached documents or web pages into the task unless they are relevant\nto the user's request. Record sources when the result will be reused.\n\nBefore editing, identify:\n\n- current behavior and compatibility contracts;\n- ownership of state and security boundaries;\n- likely failure modes, cost\u002Ftoken risks, and asynchronous behavior;\n- existing tests and the smallest relevant regression command.\n\n### 2. Make the plan executable\n\nConvert the outcome into dependency-ordered tickets. Each ticket must be\nsmall enough for one focused implementation pass and must name exact files,\ninterfaces, tests, and a stopping condition.\n\nUse this ticket shape:\n\n```text\nTicket: Txx\nObjective: one observable outcome\nRead first: exact docs\u002Ffiles\u002Fcallers\nAllowed scope: exact files or directories\nDo not change: compatibility\u002Fsecurity boundaries\nImplement: ordered, concrete changes\nTests: focused command and regression command\nAcceptance: observable pass conditions\nStop if: ambiguity, missing authority, unrelated failure, or contract conflict\nReport: files, tests, assumptions, risks, and next ticket\n```\n\nDo not create speculative tickets. Mark production-only work separately from\nlocal\u002Fdemo work. Track assumptions in one ledger with an owner and the evidence\nneeded to close them.\n\n### 3. Implement one ticket\n\nRe-read the target files and callers immediately before editing. Make the\nsmallest compatible change. Add or update a test for every non-trivial branch,\nespecially retries, timeouts, cancellation, cache invalidation, authorization,\npartial results, and restart\u002Freconnect behavior.\n\nFor agentic systems, preserve these boundaries:\n\n- planners produce typed plans, not unrestricted nested prompts;\n- schedulers own lifecycle, idempotency, leases, budgets, and terminal state;\n- tools\u002Fskills receive least-privilege context and capabilities;\n- exact evidence and artifacts remain source-linked and access-checked;\n- model output cannot bypass deterministic schema, security, quality, or\n  artifact-integrity gates.\n\n### 4. Test the ticket\n\nRun the smallest focused test first. Then run the relevant package suite and,\nwhen the change crosses a shared boundary, the full regression suite. Also run\nthe project's formatter\u002Ftype checker\u002Fcompile check when available and always\nrun `git diff --check`.\n\nTreat warnings separately from failures. Investigate flaky failures instead\nof weakening the assertion. If a test reveals a race or contract defect in a\nshared boundary, fix the boundary and add a regression test rather than\nmasking the symptom in the new ticket.\n\n### 5. Check what did not break\n\nAfter tests pass, inspect:\n\n- `git diff --stat` and the complete changed-file list;\n- API\u002FMCP\u002FCLI schemas and backward-compatible fields;\n- security, authorization, secret, and data-retention paths;\n- retry\u002Fcancellation\u002Fwait behavior and duplicate-work protection;\n- generated artifacts or rendered output when layout matters;\n- documentation, assumptions, environment examples, and ticket status.\n\nNever claim production readiness from local tests alone. State exactly what\nwas tested, what remains untested, and whether external services or\nproduction-like infrastructure are still required.\n\n## End-of-turn report\n\nAlways report:\n\n1. outcome and ticket completed;\n2. files changed;\n3. focused and regression test results;\n4. assumptions or blockers discovered;\n5. next executable ticket;\n6. explicit statement that no commit\u002Fpush was performed unless authorized.\n\nWhen a user says \"continue\", resume the first unfinished dependency in the\nactive plan, verify the working tree and remote state first, and do not repeat\ncompleted tickets. If a remote change has arrived, compare changed file sets\nand merge only with explicit user authorization; never overwrite local work.\n",{"name":5,"description":7},"main","83a1e7ce06d2399e137882bfde086cd594eb7c1a","plugins\u002Fagentic-engineering-loop\u002Fskills\u002Fagentic-engineering-loop",0,"https:\u002F\u002Fgithub.com\u002Fudit79\u002FSudarshan_agentic_core","https:\u002F\u002Fgithub.com\u002Fudit79","2026-09-23 23:27:21",{"source":8,"skillName":5,"commands":31,"prompt":41,"zipAvailable":42},[32,35,38],{"tool":33,"command":34},"npx","npx skills add udit79\u002FSudarshan_agentic_core --skill agentic-engineering-loop",{"tool":36,"command":37},"bunx","bunx skills add udit79\u002FSudarshan_agentic_core --skill agentic-engineering-loop",{"tool":39,"command":40},"pnpm","pnpm dlx skills add udit79\u002FSudarshan_agentic_core --skill agentic-engineering-loop","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fudit79\u002FSudarshan_agentic_core\u002Fblob\u002Fmain\u002Fplugins\u002Fagentic-engineering-loop\u002Fskills\u002Fagentic-engineering-loop\u002FSKILL.md",true,[44],{"path":45,"resourceType":46,"resourceTypeDesc":47,"sizeBytes":48,"sourceHtmlUrl":49},"agents\u002Fopenai.yaml","REFERENCE","参考文档",320,"https:\u002F\u002Fgithub.com\u002Fudit79\u002FSudarshan_agentic_core\u002Fblob\u002Fmain\u002Fplugins\u002Fagentic-engineering-loop\u002Fskills\u002Fagentic-engineering-loop\u002Fagents\u002Fopenai.yaml",[51,57,63,69,75,81],{"skillNo":52,"name":53,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":54,"categories":55,"keywords":56,"publishedAt":18},"01M7C1KBHTEW1BC57MHDAY99EE","visual-flowchart","https:\u002F\u002Fgithub.com\u002Fudit79\u002FSudarshan_agentic_core\u002Fblob\u002Fmain\u002Fskills\u002Fvisual-flowchart\u002FSKILL.md",[],[],{"skillNo":58,"name":59,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":60,"categories":61,"keywords":62,"publishedAt":18},"01M7C1K946K54YQ15AWMWK6A5E","video-storyboard","https:\u002F\u002Fgithub.com\u002Fudit79\u002FSudarshan_agentic_core\u002Fblob\u002Fmain\u002Fskills\u002Fvideo-storyboard\u002FSKILL.md",[],[],{"skillNo":64,"name":65,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":18},"01M7C1K61D763ZGR1QTE7V6CTT","presentation-case-brief","https:\u002F\u002Fgithub.com\u002Fudit79\u002FSudarshan_agentic_core\u002Fblob\u002Fmain\u002Fskills\u002Fpresentation-case-brief\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":72,"categories":73,"keywords":74,"publishedAt":18},"01M7C1K376TBG135GCVPM8W8MS","linkedin-reshare","https:\u002F\u002Fgithub.com\u002Fudit79\u002FSudarshan_agentic_core\u002Fblob\u002Fmain\u002Fskills\u002Flinkedin-reshare\u002FSKILL.md",[],[],{"skillNo":76,"name":77,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":78,"categories":79,"keywords":80,"publishedAt":18},"01M7C1K21M13VB7H9DVAKDYRWA","linkedin-reply","https:\u002F\u002Fgithub.com\u002Fudit79\u002FSudarshan_agentic_core\u002Fblob\u002Fmain\u002Fskills\u002Flinkedin-reply\u002FSKILL.md",[],[],{"skillNo":82,"name":83,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":6,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":18},"01M7C1K12K6K95AT4761472JKD","linkedin-post","https:\u002F\u002Fgithub.com\u002Fudit79\u002FSudarshan_agentic_core\u002Fblob\u002Fmain\u002Fskills\u002Flinkedin-post\u002FSKILL.md",[],[]]