[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6ACM7ZPQ9KDPE7KMHM6TWQ3":3,"skill-same:01M6ACM7ZPQ9KDPE7KMHM6TWQ3":66},{"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":23,"publishedAt":32,"linkOnly":33,"markdownContent":34,"markdownBody":35,"frontMatter":36,"refName":37,"commitSha":38,"skillDirectory":39,"forksCount":40,"repositoryHtmlUrl":41,"repositoryHomepageUrl":42,"ownerHtmlUrl":43,"ownerBlogUrl":44,"ownerDescription":6,"updatedAt":45,"install":46,"resources":59},"01M6ACM7ZPQ9KDPE7KMHM6TWQ3","pixiu-llm-gateway",null,"Creates and validates dubbo-go-pixiu LLM gateway conf.yaml. Use for LLM proxy\u002Ftokenizer\u002Fkvcache filters, llm_meta, vLLM, LMCache, retry\u002Ffallback, or Nacos LLM discovery. Do not use for MCP gateway config.","apache\u002Fdubbo-go-pixiu","apache","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F47359?v=4","01M6AC83E6K4FDKK7HZ8WNMQKE","The Apache Software Foundation","01M6ACKYWD1D4GF59J6YVD0YXK",568,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fapache\u002Fdubbo-go-pixiu\u002Fblob\u002Fdevelop\u002F.agents\u002Fskills\u002Fpixiu-llm-gateway\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发","SCENARIO","使用场景",[24,25,26,27,28,29,30,31],"gateway","pixiu llm gateway","MCP","go","config","mcp gateway","dubbo","proxy","2026-11-10 10:12:30",false,"---\nname: pixiu-llm-gateway\ndescription: Creates and validates dubbo-go-pixiu LLM gateway conf.yaml. Use for LLM proxy\u002Ftokenizer\u002Fkvcache filters, llm_meta, vLLM, LMCache, retry\u002Ffallback, or Nacos LLM discovery. Do not use for MCP gateway config.\n---\n\n# pixiu-llm-gateway\n\n## Purpose\nGenerate a complete or embeddable LLM gateway `conf.yaml` that configures Pixiu as a multi-provider HTTP proxy.\n\n## When to use\n- Use when:\n  - Generating or fixing Pixiu LLM gateway `conf.yaml`, including LLM proxy\u002Ftokenizer\u002Fkvcache, `llm_meta`, retry\u002Ffallback, vLLM\u002FLMCache, or Nacos LLM discovery.\n\n- Do not use for:\n  - MCP gateway, HTTP-to-Dubbo routes, or implementing new LLM-related filters.\n\n## Inputs\n\n\u003CHARD-GATE>\nDo not generate YAML, write code, create files, or take any implementation action until the user has provided all required inputs. This is a first principle.\n\nIf any required input is missing, this turn must only ask for the missing fields in the current input group; do not generate examples, defaults, YAML, code, or final output.\n\nEven if configuration information seems inferable, obvious, or implied by context, you must still ask the user to confirm it. Do not proceed until the user confirms it.\n\u003C\u002FHARD-GATE>\n\n- Choose upstream mode (required):\n  - Required:\n    - `upstream_mode`: `static` or `registry`\n\n- `listener` (required):\n  - Defaults:\n    - `address.socket_address.address: 0.0.0.0`\n    - `address.socket_address.port: 8888`\n\n- `route_config` (required):\n  - Defaults:\n    - `routes[].match.prefix: \u002Fv1`\n    - `routes[].route.cluster: llm`\n\n- `dgp.filter.llm.proxy` (required):\n  - Defaults:\n    - `config.scheme: http` (use `https` when the upstream is an HTTPS endpoint)\n    - `config.timeout: 60s`\n    - `config.maxIdleConns: 100`\n    - `config.maxIdleConnsPerHost: 100`\n    - `config.maxConnsPerHost: 100`\n\n- Static upstream (required when `upstream_mode: static`):\n  - Required:\n    - `endpoints[].ID`\n    - `endpoints[].socket_address.address` or `endpoints[].socket_address.domains`\n    - `endpoints[].socket_address.port`\n  - Optional:\n    - `endpoints[].llm_meta.provider`\n    - `endpoints[].llm_meta.api_key`\n    - `endpoints[].llm_meta.retry_policy.config`\n  - Defaults:\n    - `clusters[].name: llm`\n    - `clusters[].lb_policy: RoundRobin`\n    - `endpoints[].llm_meta.retry_policy.name: NoRetry`\n    - `endpoints[].llm_meta.fallback: false`\n    - `endpoints[].llm_meta.health_check_interval: 5000`\n\n- Registry upstream (required when `upstream_mode: registry`):\n  - Required:\n    - `registries.nacos.address`\n  - Optional:\n    - `registries.nacos.group`\n    - `registries.nacos.namespace`\n    - `registries.nacos.username`\n    - `registries.nacos.password`\n  - Defaults:\n    - `adapters[].id: llm-registry`\n    - `adapters[].name: dgp.adapter.llmregistrycenter`\n    - `registries.nacos.protocol: nacos`\n    - `registries.nacos.timeout: 5s`\n    - `registries.nacos.group: DEFAULT_GROUP`\n\n- `dgp.filter.llm.tokenizer` (optional):\n  - Defaults:\n    - `config.log_to_console: false`\n\n- `dgp.filter.ai.kvcache` (optional):\n  - Required:\n    - `config.enabled: true` (must be set explicitly when enabling kvcache)\n    - `config.vllm_endpoint`\n    - `config.lmcache_endpoint`\n  - Optional:\n    - `config.default_model`\n    - `config.token_cache`\n    - `config.cache_strategy`\n  - Defaults:\n    - `config.request_timeout: 2s`\n    - `config.lookup_routing_timeout: 50ms`\n    - `config.hot_window: 5m`\n    - `config.hot_max_records: 300`\n    - `config.max_idle_conns: 100`\n    - `config.max_idle_conns_per_host: 100`\n    - `config.max_conns_per_host: 100`\n    - `config.retry.max_attempts: 3`\n    - `config.retry.base_backoff: 100ms`\n    - `config.retry.max_backoff: 2s`\n    - `config.circuit_breaker.failure_threshold: 5`\n    - `config.circuit_breaker.recovery_timeout: 10s`\n    - `config.circuit_breaker.half_open_max_calls: 2`\n\n## Workflow\n1. Check required Inputs first:\n   1. Read existing config and already provided user information first; do not ask again for information already present or stated.\n   2. Ask one `Inputs` group at a time. Each time, output only that group's required fields, with a short explanation after each field.\n   3. If the current group's required fields are incomplete, ask only for the missing fields and do not move to the next group.\n   4. After the current group's required fields are complete, ask whether to fill that group's optional fields; if yes, list those optional fields with short explanations.\n   5. After optional fields are skipped or completed, apply that group's defaults and output default-value information; defaults must not override existing config or user input.\n2. Read current source before generating YAML:\n   - `pkg\u002Fcommon\u002Fconstant\u002Fkey.go`.\n   - `pkg\u002Ffilter\u002Fllm\u002Fproxy\u002Ffilter.go`.\n   - `pkg\u002Ffilter\u002Fllm\u002Ftokenizer\u002Ftokenizer.go`.\n   - `pkg\u002Ffilter\u002Fai\u002Fkvcache\u002Fconfig.go` and `pkg\u002Ffilter\u002Fai\u002Fkvcache\u002Fhandlers.go`.\n   - `pkg\u002Fmodel\u002Fllm.go`, `pkg\u002Fmodel\u002Fcluster.go`, and `pkg\u002Fmodel\u002Fbase.go`.\n3. Choose static or registry path by `upstream_mode`: static uses `static_resources.clusters[]`; registry uses the LLM registry adapter.\n4. Generate listener, route, and filters: use HCM to carry the LLM route and HTTP filters; add `dgp.filter.ai.kvcache`, tokenizer, and `dgp.filter.llm.proxy` as needed.\n5. Generate LLM upstream: static endpoints go under `socket_address` and `llm_meta`; ensure the LLM cluster does not mix in ordinary HTTP endpoints.\n\n## Output format\n- Show the relevant YAML fragments.\n\n## Validation\n- Verify LLM filter-chain order: `dgp.filter.ai.kvcache` -> `dgp.filter.llm.tokenizer` -> `dgp.filter.llm.proxy`; ignore missing filters within this order chain.\n- Verify `scheme` is under `dgp.filter.llm.proxy.config`, not under an endpoint; `socket_address.domains` contains host names only, such as `api.openai.com`, not full URLs or paths.\n- When KV cache is enabled, verify the LMCache-side `instance_id` exactly matches the Pixiu endpoint `ID`.\n\n## Examples\nComplete LLM route example (`conf.yaml`):\n\n```yaml\nstatic_resources:\n  listeners:\n    - name: net\u002Fhttp\n      protocol_type: HTTP\n      address:\n        socket_address:\n          address: 0.0.0.0\n          port: 8888\n      filter_chains:\n        filters:\n          - name: dgp.filter.httpconnectionmanager\n            config:\n              route_config:\n                routes:\n                  - match:\n                      prefix: \u002Fv1\n                    route:\n                      cluster: llm\n              http_filters:\n                - name: dgp.filter.ai.kvcache\n                  config:\n                    enabled: true\n                    vllm_endpoint: \"http:\u002F\u002F127.0.0.1:8000\"\n                    lmcache_endpoint: \"http:\u002F\u002F127.0.0.1:9000\"\n                    default_model: \"Qwen2.5-3B-Instruct\"\n                    request_timeout: \"2s\"\n                    lookup_routing_timeout: \"50ms\"\n                    hot_window: \"5m\"\n                    hot_max_records: 300\n                    hot_max_keys: 1000\n                    max_idle_conns: 100\n                    max_idle_conns_per_host: 100\n                    max_conns_per_host: 100\n                    token_cache:\n                      enabled: true\n                      max_size: 1024\n                      ttl: \"10m\"\n                    cache_strategy:\n                      enable_compression: true\n                      enable_pinning: true\n                      enable_eviction: true\n                      memory_threshold: 0.85\n                      hot_content_threshold: 10\n                      load_threshold: 0.7\n                      pin_instance_id: \"vllm-instance-1\"\n                      pin_location: \"LocalCPUBackend\"\n                      compress_instance_id: \"vllm-instance-1\"\n                      compress_location: \"LocalCPUBackend\"\n                      compress_method: \"zstd\"\n                      evict_instance_id: \"vllm-instance-1\"\n                    circuit_breaker:\n                      failure_threshold: 5\n                      recovery_timeout: \"10s\"\n                      half_open_max_calls: 2\n                    retry:\n                      max_attempts: 3\n                      base_backoff: \"100ms\"\n                      max_backoff: \"2s\"\n                - name: dgp.filter.llm.tokenizer\n                  config:\n                    log_to_console: false\n                - name: dgp.filter.llm.proxy\n                  config:\n                    scheme: http\n                    timeout: \"60s\"\n                    maxIdleConns: 100\n                    maxIdleConnsPerHost: 100\n                    maxConnsPerHost: 100\n  clusters:\n    - name: llm\n      lb_policy: RoundRobin\n      endpoints:\n        - ID: vllm-instance-1\n          socket_address:\n            address: 127.0.0.1\n            port: 8000\n          llm_meta:\n            provider: vllm\n            api_key: \"\u003Capi-key>\"\n            fallback: false\n            health_check_interval: 5000\n            retry_policy:\n              name: ExponentialBackoff\n              config:\n                times: 3\n                initialInterval: \"200ms\"\n                maxInterval: \"5s\"\n                multiplier: 2.0\n```\n\nNacos LLM registry mode example (`conf.yaml`):\n\n```yaml\nstatic_resources:\n  listeners:\n    - name: net\u002Fhttp\n      protocol_type: HTTP\n      address:\n        socket_address:\n          address: 0.0.0.0\n          port: 8888\n      filter_chains:\n        filters:\n          - name: dgp.filter.httpconnectionmanager\n            config:\n              route_config:\n                routes:\n                  - match:\n                      prefix: \u002Fv1\n                    route:\n                      cluster: llm\n              http_filters:\n                - name: dgp.filter.llm.proxy\n                  config:\n                    scheme: http\n                    timeout: \"60s\"\n  adapters:\n    - id: llm-nacos\n      name: dgp.adapter.llmregistrycenter\n      config:\n        registries:\n          nacos:\n            protocol: nacos\n            address: \"127.0.0.1:8848\"\n            timeout: \"5s\"\n            group: DEFAULT_GROUP\n            namespace: public\n```\n","# pixiu-llm-gateway\n\n## Purpose\nGenerate a complete or embeddable LLM gateway `conf.yaml` that configures Pixiu as a multi-provider HTTP proxy.\n\n## When to use\n- Use when:\n  - Generating or fixing Pixiu LLM gateway `conf.yaml`, including LLM proxy\u002Ftokenizer\u002Fkvcache, `llm_meta`, retry\u002Ffallback, vLLM\u002FLMCache, or Nacos LLM discovery.\n\n- Do not use for:\n  - MCP gateway, HTTP-to-Dubbo routes, or implementing new LLM-related filters.\n\n## Inputs\n\n\u003CHARD-GATE>\nDo not generate YAML, write code, create files, or take any implementation action until the user has provided all required inputs. This is a first principle.\n\nIf any required input is missing, this turn must only ask for the missing fields in the current input group; do not generate examples, defaults, YAML, code, or final output.\n\nEven if configuration information seems inferable, obvious, or implied by context, you must still ask the user to confirm it. Do not proceed until the user confirms it.\n\u003C\u002FHARD-GATE>\n\n- Choose upstream mode (required):\n  - Required:\n    - `upstream_mode`: `static` or `registry`\n\n- `listener` (required):\n  - Defaults:\n    - `address.socket_address.address: 0.0.0.0`\n    - `address.socket_address.port: 8888`\n\n- `route_config` (required):\n  - Defaults:\n    - `routes[].match.prefix: \u002Fv1`\n    - `routes[].route.cluster: llm`\n\n- `dgp.filter.llm.proxy` (required):\n  - Defaults:\n    - `config.scheme: http` (use `https` when the upstream is an HTTPS endpoint)\n    - `config.timeout: 60s`\n    - `config.maxIdleConns: 100`\n    - `config.maxIdleConnsPerHost: 100`\n    - `config.maxConnsPerHost: 100`\n\n- Static upstream (required when `upstream_mode: static`):\n  - Required:\n    - `endpoints[].ID`\n    - `endpoints[].socket_address.address` or `endpoints[].socket_address.domains`\n    - `endpoints[].socket_address.port`\n  - Optional:\n    - `endpoints[].llm_meta.provider`\n    - `endpoints[].llm_meta.api_key`\n    - `endpoints[].llm_meta.retry_policy.config`\n  - Defaults:\n    - `clusters[].name: llm`\n    - `clusters[].lb_policy: RoundRobin`\n    - `endpoints[].llm_meta.retry_policy.name: NoRetry`\n    - `endpoints[].llm_meta.fallback: false`\n    - `endpoints[].llm_meta.health_check_interval: 5000`\n\n- Registry upstream (required when `upstream_mode: registry`):\n  - Required:\n    - `registries.nacos.address`\n  - Optional:\n    - `registries.nacos.group`\n    - `registries.nacos.namespace`\n    - `registries.nacos.username`\n    - `registries.nacos.password`\n  - Defaults:\n    - `adapters[].id: llm-registry`\n    - `adapters[].name: dgp.adapter.llmregistrycenter`\n    - `registries.nacos.protocol: nacos`\n    - `registries.nacos.timeout: 5s`\n    - `registries.nacos.group: DEFAULT_GROUP`\n\n- `dgp.filter.llm.tokenizer` (optional):\n  - Defaults:\n    - `config.log_to_console: false`\n\n- `dgp.filter.ai.kvcache` (optional):\n  - Required:\n    - `config.enabled: true` (must be set explicitly when enabling kvcache)\n    - `config.vllm_endpoint`\n    - `config.lmcache_endpoint`\n  - Optional:\n    - `config.default_model`\n    - `config.token_cache`\n    - `config.cache_strategy`\n  - Defaults:\n    - `config.request_timeout: 2s`\n    - `config.lookup_routing_timeout: 50ms`\n    - `config.hot_window: 5m`\n    - `config.hot_max_records: 300`\n    - `config.max_idle_conns: 100`\n    - `config.max_idle_conns_per_host: 100`\n    - `config.max_conns_per_host: 100`\n    - `config.retry.max_attempts: 3`\n    - `config.retry.base_backoff: 100ms`\n    - `config.retry.max_backoff: 2s`\n    - `config.circuit_breaker.failure_threshold: 5`\n    - `config.circuit_breaker.recovery_timeout: 10s`\n    - `config.circuit_breaker.half_open_max_calls: 2`\n\n## Workflow\n1. Check required Inputs first:\n   1. Read existing config and already provided user information first; do not ask again for information already present or stated.\n   2. Ask one `Inputs` group at a time. Each time, output only that group's required fields, with a short explanation after each field.\n   3. If the current group's required fields are incomplete, ask only for the missing fields and do not move to the next group.\n   4. After the current group's required fields are complete, ask whether to fill that group's optional fields; if yes, list those optional fields with short explanations.\n   5. After optional fields are skipped or completed, apply that group's defaults and output default-value information; defaults must not override existing config or user input.\n2. Read current source before generating YAML:\n   - `pkg\u002Fcommon\u002Fconstant\u002Fkey.go`.\n   - `pkg\u002Ffilter\u002Fllm\u002Fproxy\u002Ffilter.go`.\n   - `pkg\u002Ffilter\u002Fllm\u002Ftokenizer\u002Ftokenizer.go`.\n   - `pkg\u002Ffilter\u002Fai\u002Fkvcache\u002Fconfig.go` and `pkg\u002Ffilter\u002Fai\u002Fkvcache\u002Fhandlers.go`.\n   - `pkg\u002Fmodel\u002Fllm.go`, `pkg\u002Fmodel\u002Fcluster.go`, and `pkg\u002Fmodel\u002Fbase.go`.\n3. Choose static or registry path by `upstream_mode`: static uses `static_resources.clusters[]`; registry uses the LLM registry adapter.\n4. Generate listener, route, and filters: use HCM to carry the LLM route and HTTP filters; add `dgp.filter.ai.kvcache`, tokenizer, and `dgp.filter.llm.proxy` as needed.\n5. Generate LLM upstream: static endpoints go under `socket_address` and `llm_meta`; ensure the LLM cluster does not mix in ordinary HTTP endpoints.\n\n## Output format\n- Show the relevant YAML fragments.\n\n## Validation\n- Verify LLM filter-chain order: `dgp.filter.ai.kvcache` -> `dgp.filter.llm.tokenizer` -> `dgp.filter.llm.proxy`; ignore missing filters within this order chain.\n- Verify `scheme` is under `dgp.filter.llm.proxy.config`, not under an endpoint; `socket_address.domains` contains host names only, such as `api.openai.com`, not full URLs or paths.\n- When KV cache is enabled, verify the LMCache-side `instance_id` exactly matches the Pixiu endpoint `ID`.\n\n## Examples\nComplete LLM route example (`conf.yaml`):\n\n```yaml\nstatic_resources:\n  listeners:\n    - name: net\u002Fhttp\n      protocol_type: HTTP\n      address:\n        socket_address:\n          address: 0.0.0.0\n          port: 8888\n      filter_chains:\n        filters:\n          - name: dgp.filter.httpconnectionmanager\n            config:\n              route_config:\n                routes:\n                  - match:\n                      prefix: \u002Fv1\n                    route:\n                      cluster: llm\n              http_filters:\n                - name: dgp.filter.ai.kvcache\n                  config:\n                    enabled: true\n                    vllm_endpoint: \"http:\u002F\u002F127.0.0.1:8000\"\n                    lmcache_endpoint: \"http:\u002F\u002F127.0.0.1:9000\"\n                    default_model: \"Qwen2.5-3B-Instruct\"\n                    request_timeout: \"2s\"\n                    lookup_routing_timeout: \"50ms\"\n                    hot_window: \"5m\"\n                    hot_max_records: 300\n                    hot_max_keys: 1000\n                    max_idle_conns: 100\n                    max_idle_conns_per_host: 100\n                    max_conns_per_host: 100\n                    token_cache:\n                      enabled: true\n                      max_size: 1024\n                      ttl: \"10m\"\n                    cache_strategy:\n                      enable_compression: true\n                      enable_pinning: true\n                      enable_eviction: true\n                      memory_threshold: 0.85\n                      hot_content_threshold: 10\n                      load_threshold: 0.7\n                      pin_instance_id: \"vllm-instance-1\"\n                      pin_location: \"LocalCPUBackend\"\n                      compress_instance_id: \"vllm-instance-1\"\n                      compress_location: \"LocalCPUBackend\"\n                      compress_method: \"zstd\"\n                      evict_instance_id: \"vllm-instance-1\"\n                    circuit_breaker:\n                      failure_threshold: 5\n                      recovery_timeout: \"10s\"\n                      half_open_max_calls: 2\n                    retry:\n                      max_attempts: 3\n                      base_backoff: \"100ms\"\n                      max_backoff: \"2s\"\n                - name: dgp.filter.llm.tokenizer\n                  config:\n                    log_to_console: false\n                - name: dgp.filter.llm.proxy\n                  config:\n                    scheme: http\n                    timeout: \"60s\"\n                    maxIdleConns: 100\n                    maxIdleConnsPerHost: 100\n                    maxConnsPerHost: 100\n  clusters:\n    - name: llm\n      lb_policy: RoundRobin\n      endpoints:\n        - ID: vllm-instance-1\n          socket_address:\n            address: 127.0.0.1\n            port: 8000\n          llm_meta:\n            provider: vllm\n            api_key: \"\u003Capi-key>\"\n            fallback: false\n            health_check_interval: 5000\n            retry_policy:\n              name: ExponentialBackoff\n              config:\n                times: 3\n                initialInterval: \"200ms\"\n                maxInterval: \"5s\"\n                multiplier: 2.0\n```\n\nNacos LLM registry mode example (`conf.yaml`):\n\n```yaml\nstatic_resources:\n  listeners:\n    - name: net\u002Fhttp\n      protocol_type: HTTP\n      address:\n        socket_address:\n          address: 0.0.0.0\n          port: 8888\n      filter_chains:\n        filters:\n          - name: dgp.filter.httpconnectionmanager\n            config:\n              route_config:\n                routes:\n                  - match:\n                      prefix: \u002Fv1\n                    route:\n                      cluster: llm\n              http_filters:\n                - name: dgp.filter.llm.proxy\n                  config:\n                    scheme: http\n                    timeout: \"60s\"\n  adapters:\n    - id: llm-nacos\n      name: dgp.adapter.llmregistrycenter\n      config:\n        registries:\n          nacos:\n            protocol: nacos\n            address: \"127.0.0.1:8848\"\n            timeout: \"5s\"\n            group: DEFAULT_GROUP\n            namespace: public\n```\n",{"name":5,"description":7},"develop","42c72431c93dbd5fa7939fb3583ba56b8ef6ceaf",".agents\u002Fskills\u002Fpixiu-llm-gateway",195,"https:\u002F\u002Fgithub.com\u002Fapache\u002Fdubbo-go-pixiu","https:\u002F\u002Fdubbogo.github.io\u002Fdubbo-go-proxy\u002F","https:\u002F\u002Fgithub.com\u002Fapache","https:\u002F\u002Fwww.apache.org\u002F","2026-11-13 17:51:26",{"source":8,"skillName":5,"commands":47,"prompt":57,"zipAvailable":58},[48,51,54],{"tool":49,"command":50},"npx","npx skills add apache\u002Fdubbo-go-pixiu --skill pixiu-llm-gateway",{"tool":52,"command":53},"bunx","bunx skills add apache\u002Fdubbo-go-pixiu --skill pixiu-llm-gateway",{"tool":55,"command":56},"pnpm","pnpm dlx skills add apache\u002Fdubbo-go-pixiu --skill pixiu-llm-gateway","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fapache\u002Fdubbo-go-pixiu\u002Fblob\u002Fdevelop\u002F.agents\u002Fskills\u002Fpixiu-llm-gateway\u002FSKILL.md",true,[60],{"path":61,"resourceType":62,"resourceTypeDesc":63,"sizeBytes":64,"sourceHtmlUrl":65},"agents\u002Fopenai.yaml","REFERENCE","参考文档",141,"https:\u002F\u002Fgithub.com\u002Ffmal\u002Fdotfiles\u002Fblob\u002Fmaster\u002F.agents\u002Fskills\u002Fclaude-handoff\u002Fagents\u002Fopenai.yaml",[67,84,92],{"skillNo":68,"name":69,"nameZh":6,"description":70,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":71,"categories":72,"keywords":77,"publishedAt":32},"01M6ACMAM1XGCW12W3XXGARSC9","pixiu-mcp-gateway","Creates and validates dubbo-go-pixiu MCP gateway conf.yaml. Use for MCP server filters, tools\u002Flist, tools\u002Fcall, Mcp-Session-Id, MCP auth, Nacos MCP registry, or exposing backend HTTP APIs as MCP tools. Do not use for LLM model proxying.","https:\u002F\u002Fgithub.com\u002Fapache\u002Fdubbo-go-pixiu\u002Fblob\u002Fdevelop\u002F.agents\u002Fskills\u002Fpixiu-mcp-gateway\u002FSKILL.md",[73,76],{"categoryNo":74,"categoryName":75,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000010","后端与接口",{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[78,24,29,26,27,79,80,81,82,83,30],"pixiu mcp gateway","mcp server","auth","http","registry","mcp tools",{"skillNo":85,"name":86,"nameZh":6,"description":87,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":88,"categories":89,"keywords":90,"publishedAt":32},"01M6ACM57QRS30NCDCVBRMDGNK","pixiu-http-to-dubbo","Creates and debugs dubbo-go-pixiu HTTP-to-Dubbo route YAML. Use for api_config.yaml\u002Fconf.yaml routes, direct Dubbo URLs, registry-backed Dubbo, mappingParams, mapType, parameterTypes, opt.types\u002Fopt.values, or generic-invoke errors. Do not use for non-Dubbo HTTP proxy routes or new registry adapter implementation.","https:\u002F\u002Fgithub.com\u002Fapache\u002Fdubbo-go-pixiu\u002Fblob\u002Fdevelop\u002F.agents\u002Fskills\u002Fpixiu-http-to-dubbo\u002FSKILL.md",[],[91,30,81,28,82,27,31],"pixiu http to dubbo",{"skillNo":93,"name":94,"nameZh":6,"description":95,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":96,"categories":97,"keywords":98,"publishedAt":32},"01M6ACM0RKAZJYS1NA0M15ZB1N","pixiu-filter-author","Creates and debugs dubbo-go-pixiu HTTP\u002FNetwork filters. Use when adding, registering, or wiring Pixiu filters, or when a filter compiles but does not run. Do not use for adapter or protocol-listener SPI work.","https:\u002F\u002Fgithub.com\u002Fapache\u002Fdubbo-go-pixiu\u002Fblob\u002Fdevelop\u002F.agents\u002Fskills\u002Fpixiu-filter-author\u002FSKILL.md",[],[99,27,100,101,81],"pixiu filter author","protocol","run"]