[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7D2JSWWDGZBENQ96RCKXVNV":3,"skill-same:01M7D2JSWWDGZBENQ96RCKXVNV":44},{"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":14,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":29,"install":30,"resources":43},"01M7D2JSWWDGZBENQ96RCKXVNV","backend-integration-testing",null,"Enforces Skillscope's pytest integration-test conventions for the Python backend. Use when testing SQLite repositories and migrations, Cursor transcript and hook adapters, FastAPI routes, CLI wiring, composition roots, filesystem fixtures, or code under tests\u002Fintegration.","ganan-s\u002Fskillscope","ganan-s","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F35375634?v=4","01M7D2JD9JNTDC32EMNQK3SZ8K","Ganan","01M7D2JDA79T5R2JXDV9RWEAG3",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fganan-s\u002Fskillscope\u002Fblob\u002Fmain\u002F.cursor\u002Fskills\u002Fbackend-integration-testing\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: backend-integration-testing\ndescription: Enforces Skillscope's pytest integration-test conventions for the Python backend. Use when testing SQLite repositories and migrations, Cursor transcript and hook adapters, FastAPI routes, CLI wiring, composition roots, filesystem fixtures, or code under tests\u002Fintegration.\n---\n\n# Backend integration testing\n\nApply these conventions to Skillscope backend integration tests. Also apply\n`backend-hexagonal-architecture` when a test exposes or changes an architectural\nboundary.\n\n## Role in the testing pyramid\n\nIntegration tests form the middle of the test pyramid. They prove that two or\nmore real Skillscope components honor their shared contract.\n\nAn integration test may use:\n\n- real SQLite connections and production migrations;\n- real temporary files and directories;\n- sanitized Cursor transcript and hook-spool fixtures;\n- the FastAPI application through an in-process test client;\n- the real composition root with test configuration; and\n- an in-process CLI runner.\n\nAn integration test must not use:\n\n- the user's actual home directory, Cursor data, hooks, or Skillscope database;\n- internet or external service access;\n- a separately running backend process;\n- a browser;\n- arbitrary timing waits; or\n- test-order dependencies.\n\nWorkflows requiring an installed application, real process boundaries, or a\nbrowser belong in the end-to-end suite.\n\n## Framework and location\n\nUse `pytest`.\n\nPlace tests under `tests\u002Fintegration\u002F`, organized by boundary:\n\n```text\ntests\u002Fintegration\u002F\n  api\u002F\n  cli\u002F\n  composition\u002F\n  plugins\u002F\n    cursor\u002F\n  storage\u002F\n  fixtures\u002F\n    cursor\u002F\n```\n\nName files `test_\u003Cboundary>.py`. Name tests:\n\n```text\ntest_\u003Coperation>_\u003Ccondition>_\u003Cobservable_outcome>\n```\n\nPrefer test functions. Group scenarios in classes only when that improves\nnavigation; do not use test-class inheritance.\n\n## What to integration test\n\nPrioritize contracts that unit tests deliberately replace with ports or\nin-memory values:\n\n- ordered SQL migrations against an empty and previously migrated database;\n- SQLite repository reads, writes, transactions, constraints, and row mapping;\n- read-only serve connections and write rejection;\n- Cursor transcript discovery and parsing from filesystem fixtures;\n- Cursor hook-spool parsing and merge precedence;\n- application use cases with real SQLite adapters;\n- FastAPI routing, validation, error mapping, OpenAPI, and JSON serialization;\n- production dependency wiring with test configuration;\n- CLI option parsing and exit behavior; and\n- compiled frontend mount\u002Ffallback behavior when frontend assets exist.\n\nDo not duplicate pure domain cases already covered by unit tests. Exercise the\nintegration seam and its important failure modes.\n\n## Test boundaries\n\nEach test must state which real boundary it proves. Common scopes are:\n\n### Storage adapter\n\nUse the real connection factory, migrations, repositories, and explicit SQL.\nThe application service may be real when transaction ownership is part of the\ncontract.\n\n### Cursor adapter\n\nUse real fixture files and the real Cursor plugin. Inject the transcript root,\nhook-spool path, platform, and clock. Never scan `~\u002F.cursor`.\n\n### REST API\n\nBuild the real FastAPI application with real query services and a temporary\nsnapshot database. Use an in-process client. Do not mock route handlers or\nFastAPI internals.\n\n### CLI\n\nInvoke the real command registration and composition through an in-process CLI\nrunner with explicit temporary paths. Use a subprocess only when process\nbehavior itself is the requirement; otherwise reserve that for end-to-end.\n\n### Composition\n\nConstruct the application from production wiring with test configuration and\nreplace only boundaries outside the scope of the test.\n\n## Database conventions\n\nCreate databases under `tmp_path`. Never use an in-memory SQLite database when\nthe behavior depends on:\n\n- WAL;\n- read-only URI mode;\n- multiple connections;\n- locking;\n- filesystem permissions; or\n- persistence across application instances.\n\nApply production migrations rather than creating tables directly in test code.\n\nDefault to a fresh database per test. A broader fixture scope is allowed only\nfor immutable migrated templates copied into each test.\n\nTests must cover:\n\n- migration from an empty store;\n- repeated migration as a no-op;\n- rejection of unsupported newer schemas;\n- foreign-key enforcement;\n- complete aggregate replacement in one transaction;\n- rollback without partial child rows;\n- unchanged revision idempotency;\n- stale revision protection;\n- deterministic list\u002Fdetail ordering; and\n- read-only serving that cannot create, migrate, or mutate the store.\n\nAssert through repository or query-port results. Direct SQL assertions are\nappropriate only when verifying schema, constraints, or atomicity that the\npublic port cannot expose.\n\n## Cursor fixture conventions\n\nStore only synthetic or sanitized native fixtures under\n`tests\u002Fintegration\u002Ffixtures\u002Fcursor\u002F`.\n\nInclude focused fixtures for:\n\n- a closed transcript with no skills;\n- a successful `postToolUse` read of exact `SKILL.md`;\n- repeated successful activations;\n- `postToolUseFailure`;\n- an offline read request with unknown outcome;\n- malformed or truncated JSONL;\n- an active or grace-period transcript;\n- a continued conversation revision;\n- missing and malformed frontmatter;\n- resource reads before and after activation; and\n- Linux and macOS path examples.\n\nFixtures must not contain real prompts, emails, access tokens, home paths, model\nthoughts, or unrelated tool output.\n\nKeep raw native shapes in fixtures. Do not pre-normalize them into the expected\ncanonical DTO, or the test will bypass the adapter behavior it claims to prove.\n\n## Filesystem isolation\n\nAll writable paths come from `tmp_path` or an explicit test configuration.\nTests may copy read-only fixture trees into that temporary root.\n\nNever depend on:\n\n- the repository's current working directory unless that is the tested input;\n- files created by another test;\n- platform-specific `\u002Ftmp` contents;\n- environment variables from the developer's shell; or\n- absolute paths outside the test sandbox.\n\nUse `monkeypatch` only to set a process boundary that production code genuinely\nreads, such as a documented environment variable. Prefer passing explicit\nconfiguration.\n\n## FastAPI contract tests\n\nUse the real application factory and an in-process test client.\n\nVerify:\n\n- `\u002Fapi\u002Fv1\u002Fconversations` ordering, pagination, and empty skill lists;\n- `\u002Fapi\u002Fv1\u002Fconversations\u002F{id}` task\u002Fload ordering and unknown-id response;\n- `\u002Fapi\u002Fv1\u002Fmeta` API and schema compatibility fields;\n- stable error bodies and status codes;\n- Pydantic serialization of timestamps, enums, and optional values;\n- API routes remain read-only;\n- unknown `\u002Fapi\u002F...` paths return API `404`, not SPA HTML; and\n- non-API routes use the frontend fallback only when assets are configured.\n\nSeed data through the real write-side repository or a test data builder that\nuses public storage ports. Do not insert ad hoc rows that violate production\ninvariants unless the test specifically verifies corruption handling.\n\n## Test doubles\n\nUse real implementations on both sides of the integration boundary.\n\nReplace only dependencies outside that boundary:\n\n- a fixed clock;\n- deterministic ID generation;\n- a platform detector;\n- an external scheduler; or\n- frontend assets not relevant to an API test.\n\nDo not mock SQLite, FastAPI, filesystem calls, or the Cursor adapter in a test\nclaiming to integrate them.\n\nUse narrow fakes that implement application ports. Avoid patching module\ninternals.\n\n## Assertions\n\nAssert behavior visible at the boundary:\n\n- canonical snapshots and diagnostics from a plugin;\n- persisted\u002Fqueryable aggregates from repositories;\n- transaction outcomes;\n- HTTP status, headers, and JSON bodies;\n- CLI exit codes and user-facing output; and\n- files deliberately produced by the operation.\n\nDo not assert internal helper calls or private object graphs.\n\nWhen comparing large structures, prefer explicit expected DTOs or focused\ngolden JSON files. Review golden changes as API changes; never update them\nblindly.\n\n## Determinism and concurrency\n\nInject fixed UTC times and stable IDs. Control source mtimes explicitly when\ntesting quiescence or revision ordering.\n\nNever use `sleep` to wait for:\n\n- filesystem timestamp changes;\n- grace periods;\n- SQLite locks; or\n- background work.\n\nDrive clocks and synchronization explicitly. For lock tests, coordinate known\nconnection states and use bounded timeouts.\n\nTests may run in parallel. Each test owns its database, transcript root, hook\nspool, and environment overrides.\n\n## Failure-path requirements\n\nCover failures where adapters translate infrastructure behavior into the\napplication contract:\n\n- unreadable or malformed native source;\n- failed\u002Fdenied read evidence;\n- unavailable post-read snapshot;\n- database busy\u002Fconstraint failures;\n- migration checksum mismatch;\n- missing or incompatible store on serve;\n- repository rollback; and\n- invalid API query\u002Fpath parameters.\n\nAssert stable application errors or diagnostic codes. Do not expose raw SQL,\npersonal paths, or native payloads in expected REST errors.\n\n## Running the suite\n\nThe canonical command is:\n\n```bash\npytest tests\u002Fintegration\n```\n\nIntegration tests may take longer than unit tests but must remain suitable for\nevery pull request. They must run without network access, Cursor installation,\nuser hooks, external services, or a pre-existing database.\n\n## Review checklist\n\n- The test proves a real boundary between concrete components.\n- It belongs in integration rather than unit or end-to-end.\n- Production migrations, parsers, repositories, routes, or wiring are used.\n- All state is isolated under temporary paths.\n- No personal Cursor data or machine configuration is consulted.\n- Real implementations are used inside the declared boundary.\n- Mocks or fakes exist only outside that boundary.\n- Assertions target public contracts and durable outcomes.\n- Time, IDs, paths, and ordering are deterministic.\n- Important rollback, malformed-input, and read-only cases are covered.\n- The test passes alone, in parallel, and in any suite order.\n","# Backend integration testing\n\nApply these conventions to Skillscope backend integration tests. Also apply\n`backend-hexagonal-architecture` when a test exposes or changes an architectural\nboundary.\n\n## Role in the testing pyramid\n\nIntegration tests form the middle of the test pyramid. They prove that two or\nmore real Skillscope components honor their shared contract.\n\nAn integration test may use:\n\n- real SQLite connections and production migrations;\n- real temporary files and directories;\n- sanitized Cursor transcript and hook-spool fixtures;\n- the FastAPI application through an in-process test client;\n- the real composition root with test configuration; and\n- an in-process CLI runner.\n\nAn integration test must not use:\n\n- the user's actual home directory, Cursor data, hooks, or Skillscope database;\n- internet or external service access;\n- a separately running backend process;\n- a browser;\n- arbitrary timing waits; or\n- test-order dependencies.\n\nWorkflows requiring an installed application, real process boundaries, or a\nbrowser belong in the end-to-end suite.\n\n## Framework and location\n\nUse `pytest`.\n\nPlace tests under `tests\u002Fintegration\u002F`, organized by boundary:\n\n```text\ntests\u002Fintegration\u002F\n  api\u002F\n  cli\u002F\n  composition\u002F\n  plugins\u002F\n    cursor\u002F\n  storage\u002F\n  fixtures\u002F\n    cursor\u002F\n```\n\nName files `test_\u003Cboundary>.py`. Name tests:\n\n```text\ntest_\u003Coperation>_\u003Ccondition>_\u003Cobservable_outcome>\n```\n\nPrefer test functions. Group scenarios in classes only when that improves\nnavigation; do not use test-class inheritance.\n\n## What to integration test\n\nPrioritize contracts that unit tests deliberately replace with ports or\nin-memory values:\n\n- ordered SQL migrations against an empty and previously migrated database;\n- SQLite repository reads, writes, transactions, constraints, and row mapping;\n- read-only serve connections and write rejection;\n- Cursor transcript discovery and parsing from filesystem fixtures;\n- Cursor hook-spool parsing and merge precedence;\n- application use cases with real SQLite adapters;\n- FastAPI routing, validation, error mapping, OpenAPI, and JSON serialization;\n- production dependency wiring with test configuration;\n- CLI option parsing and exit behavior; and\n- compiled frontend mount\u002Ffallback behavior when frontend assets exist.\n\nDo not duplicate pure domain cases already covered by unit tests. Exercise the\nintegration seam and its important failure modes.\n\n## Test boundaries\n\nEach test must state which real boundary it proves. Common scopes are:\n\n### Storage adapter\n\nUse the real connection factory, migrations, repositories, and explicit SQL.\nThe application service may be real when transaction ownership is part of the\ncontract.\n\n### Cursor adapter\n\nUse real fixture files and the real Cursor plugin. Inject the transcript root,\nhook-spool path, platform, and clock. Never scan `~\u002F.cursor`.\n\n### REST API\n\nBuild the real FastAPI application with real query services and a temporary\nsnapshot database. Use an in-process client. Do not mock route handlers or\nFastAPI internals.\n\n### CLI\n\nInvoke the real command registration and composition through an in-process CLI\nrunner with explicit temporary paths. Use a subprocess only when process\nbehavior itself is the requirement; otherwise reserve that for end-to-end.\n\n### Composition\n\nConstruct the application from production wiring with test configuration and\nreplace only boundaries outside the scope of the test.\n\n## Database conventions\n\nCreate databases under `tmp_path`. Never use an in-memory SQLite database when\nthe behavior depends on:\n\n- WAL;\n- read-only URI mode;\n- multiple connections;\n- locking;\n- filesystem permissions; or\n- persistence across application instances.\n\nApply production migrations rather than creating tables directly in test code.\n\nDefault to a fresh database per test. A broader fixture scope is allowed only\nfor immutable migrated templates copied into each test.\n\nTests must cover:\n\n- migration from an empty store;\n- repeated migration as a no-op;\n- rejection of unsupported newer schemas;\n- foreign-key enforcement;\n- complete aggregate replacement in one transaction;\n- rollback without partial child rows;\n- unchanged revision idempotency;\n- stale revision protection;\n- deterministic list\u002Fdetail ordering; and\n- read-only serving that cannot create, migrate, or mutate the store.\n\nAssert through repository or query-port results. Direct SQL assertions are\nappropriate only when verifying schema, constraints, or atomicity that the\npublic port cannot expose.\n\n## Cursor fixture conventions\n\nStore only synthetic or sanitized native fixtures under\n`tests\u002Fintegration\u002Ffixtures\u002Fcursor\u002F`.\n\nInclude focused fixtures for:\n\n- a closed transcript with no skills;\n- a successful `postToolUse` read of exact `SKILL.md`;\n- repeated successful activations;\n- `postToolUseFailure`;\n- an offline read request with unknown outcome;\n- malformed or truncated JSONL;\n- an active or grace-period transcript;\n- a continued conversation revision;\n- missing and malformed frontmatter;\n- resource reads before and after activation; and\n- Linux and macOS path examples.\n\nFixtures must not contain real prompts, emails, access tokens, home paths, model\nthoughts, or unrelated tool output.\n\nKeep raw native shapes in fixtures. Do not pre-normalize them into the expected\ncanonical DTO, or the test will bypass the adapter behavior it claims to prove.\n\n## Filesystem isolation\n\nAll writable paths come from `tmp_path` or an explicit test configuration.\nTests may copy read-only fixture trees into that temporary root.\n\nNever depend on:\n\n- the repository's current working directory unless that is the tested input;\n- files created by another test;\n- platform-specific `\u002Ftmp` contents;\n- environment variables from the developer's shell; or\n- absolute paths outside the test sandbox.\n\nUse `monkeypatch` only to set a process boundary that production code genuinely\nreads, such as a documented environment variable. Prefer passing explicit\nconfiguration.\n\n## FastAPI contract tests\n\nUse the real application factory and an in-process test client.\n\nVerify:\n\n- `\u002Fapi\u002Fv1\u002Fconversations` ordering, pagination, and empty skill lists;\n- `\u002Fapi\u002Fv1\u002Fconversations\u002F{id}` task\u002Fload ordering and unknown-id response;\n- `\u002Fapi\u002Fv1\u002Fmeta` API and schema compatibility fields;\n- stable error bodies and status codes;\n- Pydantic serialization of timestamps, enums, and optional values;\n- API routes remain read-only;\n- unknown `\u002Fapi\u002F...` paths return API `404`, not SPA HTML; and\n- non-API routes use the frontend fallback only when assets are configured.\n\nSeed data through the real write-side repository or a test data builder that\nuses public storage ports. Do not insert ad hoc rows that violate production\ninvariants unless the test specifically verifies corruption handling.\n\n## Test doubles\n\nUse real implementations on both sides of the integration boundary.\n\nReplace only dependencies outside that boundary:\n\n- a fixed clock;\n- deterministic ID generation;\n- a platform detector;\n- an external scheduler; or\n- frontend assets not relevant to an API test.\n\nDo not mock SQLite, FastAPI, filesystem calls, or the Cursor adapter in a test\nclaiming to integrate them.\n\nUse narrow fakes that implement application ports. Avoid patching module\ninternals.\n\n## Assertions\n\nAssert behavior visible at the boundary:\n\n- canonical snapshots and diagnostics from a plugin;\n- persisted\u002Fqueryable aggregates from repositories;\n- transaction outcomes;\n- HTTP status, headers, and JSON bodies;\n- CLI exit codes and user-facing output; and\n- files deliberately produced by the operation.\n\nDo not assert internal helper calls or private object graphs.\n\nWhen comparing large structures, prefer explicit expected DTOs or focused\ngolden JSON files. Review golden changes as API changes; never update them\nblindly.\n\n## Determinism and concurrency\n\nInject fixed UTC times and stable IDs. Control source mtimes explicitly when\ntesting quiescence or revision ordering.\n\nNever use `sleep` to wait for:\n\n- filesystem timestamp changes;\n- grace periods;\n- SQLite locks; or\n- background work.\n\nDrive clocks and synchronization explicitly. For lock tests, coordinate known\nconnection states and use bounded timeouts.\n\nTests may run in parallel. Each test owns its database, transcript root, hook\nspool, and environment overrides.\n\n## Failure-path requirements\n\nCover failures where adapters translate infrastructure behavior into the\napplication contract:\n\n- unreadable or malformed native source;\n- failed\u002Fdenied read evidence;\n- unavailable post-read snapshot;\n- database busy\u002Fconstraint failures;\n- migration checksum mismatch;\n- missing or incompatible store on serve;\n- repository rollback; and\n- invalid API query\u002Fpath parameters.\n\nAssert stable application errors or diagnostic codes. Do not expose raw SQL,\npersonal paths, or native payloads in expected REST errors.\n\n## Running the suite\n\nThe canonical command is:\n\n```bash\npytest tests\u002Fintegration\n```\n\nIntegration tests may take longer than unit tests but must remain suitable for\nevery pull request. They must run without network access, Cursor installation,\nuser hooks, external services, or a pre-existing database.\n\n## Review checklist\n\n- The test proves a real boundary between concrete components.\n- It belongs in integration rather than unit or end-to-end.\n- Production migrations, parsers, repositories, routes, or wiring are used.\n- All state is isolated under temporary paths.\n- No personal Cursor data or machine configuration is consulted.\n- Real implementations are used inside the declared boundary.\n- Mocks or fakes exist only outside that boundary.\n- Assertions target public contracts and durable outcomes.\n- Time, IDs, paths, and ordering are deterministic.\n- Important rollback, malformed-input, and read-only cases are covered.\n- The test passes alone, in parallel, and in any suite order.\n",{"name":5,"description":7},"main","38802a07d3275daa93def8f87bb59ed53b4a2ad7",".cursor\u002Fskills\u002Fbackend-integration-testing","https:\u002F\u002Fgithub.com\u002Fganan-s\u002Fskillscope","https:\u002F\u002Fgithub.com\u002Fganan-s","2026-09-20 10:11:27",{"source":8,"skillName":5,"commands":31,"prompt":41,"zipAvailable":42},[32,35,38],{"tool":33,"command":34},"npx","npx skills add ganan-s\u002Fskillscope --skill backend-integration-testing",{"tool":36,"command":37},"bunx","bunx skills add ganan-s\u002Fskillscope --skill backend-integration-testing",{"tool":39,"command":40},"pnpm","pnpm dlx skills add ganan-s\u002Fskillscope --skill backend-integration-testing","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fganan-s\u002Fskillscope\u002Fblob\u002Fmain\u002F.cursor\u002Fskills\u002Fbackend-integration-testing\u002FSKILL.md",true,[],[45,52,57,64,71],{"skillNo":46,"name":47,"nameZh":6,"description":48,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":49,"categories":50,"keywords":51,"publishedAt":19},"01M7D2KBS5S4MG82STB8SG9C6S","skillscope-probe","Benign fixture used to verify Cursor read-hook payloads.","https:\u002F\u002Fgithub.com\u002Fganan-s\u002Fskillscope\u002Fblob\u002Fmain\u002Fexperiments\u002Fcursor-hook-probe\u002Ffixtures\u002Fprobe-skill\u002FSKILL.md",[],[],{"skillNo":53,"name":47,"nameZh":6,"description":48,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":54,"categories":55,"keywords":56,"publishedAt":19},"01M7D2K91SSKZKFCFJRPWN6C56","https:\u002F\u002Fgithub.com\u002Fganan-s\u002Fskillscope\u002Fblob\u002Fmain\u002Fexperiments\u002Fcursor-golden-session\u002Ffixtures\u002Fprobe-skill\u002FSKILL.md",[],[],{"skillNo":58,"name":59,"nameZh":6,"description":60,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":61,"categories":62,"keywords":63,"publishedAt":19},"01M7D2JZ5P0TX8HMCVEPMZ7VX5","backend-unit-testing","Enforces Skillscope's pytest unit-test conventions for the Python backend. Use when creating, modifying, or reviewing unit tests, domain logic, application use cases, canonical event parsing, test doubles, fixtures, or code under tests\u002Funit.","https:\u002F\u002Fgithub.com\u002Fganan-s\u002Fskillscope\u002Fblob\u002Fmain\u002F.cursor\u002Fskills\u002Fbackend-unit-testing\u002FSKILL.md",[],[],{"skillNo":65,"name":66,"nameZh":6,"description":67,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":68,"categories":69,"keywords":70,"publishedAt":19},"01M7D2JNFN05SGAZV6DWYB2PK6","backend-hexagonal-architecture","Enforces Skillscope's hexagonal Python backend architecture. Use when creating, modifying, or reviewing backend domain models, application services, ports, FastAPI routes, CLI commands, SQLite repositories, harness plugins, composition wiring, or other code under src\u002Fskillscope.","https:\u002F\u002Fgithub.com\u002Fganan-s\u002Fskillscope\u002Fblob\u002Fmain\u002F.cursor\u002Fskills\u002Fbackend-hexagonal-architecture\u002FSKILL.md",[],[],{"skillNo":72,"name":73,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":19},"01M7D2JGZX7JCC3ZRPMAT2XHV4","backend-e2e-testing","Enforces Skillscope's end-to-end testing conventions for installed wheels, subprocesses, Testcontainers, Docker-compatible runtimes, live HTTP requests, and Playwright. Use when creating or reviewing tests under tests\u002Fe2e or testing packaged service and dashboard workflows.","https:\u002F\u002Fgithub.com\u002Fganan-s\u002Fskillscope\u002Fblob\u002Fmain\u002F.cursor\u002Fskills\u002Fbackend-e2e-testing\u002FSKILL.md",[],[]]