[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-anthropic-commerce-architecture":3,"mdc-i1qjpe-key":37,"related-org-anthropic-commerce-architecture":1020,"related-repo-anthropic-commerce-architecture":1206},{"slug":4,"name":4,"fn":5,"description":6,"org":7,"tags":12,"stars":26,"repoUrl":27,"updatedAt":28,"license":29,"forks":30,"topics":31,"repo":32,"sourceUrl":35,"mdContent":36},"commerce-architecture","design and review commerce agent architecture","How the reference commerce agents of either role are structured, covering the loop, where each rule lives, skills, the backend interface, delegates, and model fields. Load when designing or reviewing the structure of a shopping or merchant agent.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},"anthropic","Anthropic","https:\u002F\u002Fpexgzepcugksgbtrxkhf.supabase.co\u002Fstorage\u002Fv1\u002Fobject\u002Fpublic\u002Forg-logos\u002Fanthropic.png","anthropics",[13,17,20,23],{"name":14,"slug":15,"type":16},"Architecture","architecture","tag",{"name":18,"slug":19,"type":16},"E-commerce","e-commerce",{"name":21,"slug":22,"type":16},"Agents","agents",{"name":24,"slug":25,"type":16},"Engineering","engineering",1735,"https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fcommerce-agents","2026-09-04T08:00:39.152322",null,277,[],{"repoUrl":27,"stars":26,"forks":30,"topics":33,"description":34},[],"Reference blueprint for building shopping and merchant agents with Claude. Examples in retail, commerce, telecom, and entertainment included.","https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fcommerce-agents\u002Ftree\u002FHEAD\u002Fplugins\u002Fcommerce-builder\u002Fskills\u002Fcommerce-architecture","---\nname: commerce-architecture\ndescription: How the reference commerce agents of either role are structured, covering the loop, where each rule lives, skills, the backend interface, delegates, and model fields. Load when designing or reviewing the structure of a shopping or merchant agent.\n---\n\n# Commerce agent architecture\n\nPaths are in the reference repo: `commerce_common\u002F` is `commerce-common\u002Fcommerce_common\u002F`, `shopping_agent\u002F` is\n`shopping-agent\u002Fcore\u002Fshopping_agent\u002F`, `merchant_agent\u002F` is `merchant-agent\u002Fcore\u002Fmerchant_agent\u002F`, and the runtimes are\n`shopping-agent\u002Fruntime-messages-api\u002Fshopping_agent_runtime\u002F` and `merchant-agent\u002Fruntime-messages-api\u002Fmerchant_agent_runtime\u002F`.\n\n## One loop\n\n- One model owns the conversation. A turn is `ShoppingAgent.stream_turn` or `MerchantAgent.stream_turn` (each\n  runtime's `orchestrator.py`): the model reads the cached prompt, calls tools (one round's calls run\n  concurrently), and ends with text plus presentation calls. There is no router, classifier, or hand-off.\n- Every call on every path runs through the role executor (`ShoppingToolExecutor` in `shopping_agent\u002Fexecutor.py`,\n  `MerchantToolExecutor` in `merchant_agent\u002Fexecutor.py`) over `BaseToolExecutor` in `commerce_common\u002Fexecution.py`.\n  The Messages API runtime, the Agent SDK toolset, and the MCP server call the same `execute`, which never raises;\n  after `max_tool_iterations` rounds (`commerce_common\u002Fconfig.py`) the runtime forces a round without tools.\n\n## Where a rule lives\n\n| A rule that applies | Lives in | Reference |\n|---|---|---|\n| While one call's arguments are being filled in | That tool's description | `build_tools` in each role's `tools\u002Fregistry.py` |\n| On most turns: cart and checkout, the staged-write contract, presentation grammar, tool order, trust rules | The static prompt | `build_static_system` in each role's `prompt.py` |\n| On the minority of requests that need a multi-step procedure | A skill, loaded on demand | `shopping-agent\u002Fskills\u002F`, `merchant-agent\u002Fskills\u002F` |\n\nA rule the core journey needs on most conversations moves a layer down. This table is the one statement of the layering.\n\n## Skills\n\n- A skill is a directory holding `SKILL.md`: frontmatter `name` and `description`, then the body (`parse_skill_md`\n  in `commerce_common\u002Fskills.py`). The description names the request class; it carries no sample utterances.\n- The static prompt carries the index alone (`SkillRegistry.index_block`, sorted by name) and `load_skill` returns\n  one body (`_load_skill` in `commerce_common\u002Fexecution.py`). On the Agent SDK, `ensure_project_skills` links the\n  same directories into `.claude\u002Fskills\u002F` and `SKILL_TOOL_ADAPTER` points the model at the SDK's `Skill` tool\n  (`commerce_common\u002Fagent_sdk.py`); the hosted manifests list the directories under `skills[]` in `agent.yaml`.\n\n## The backend interface\n\nA deployment implements one abstract class per role; nothing else reaches its systems.\n\n- `StorefrontBackend` (`shopping_agent\u002Fbackend.py`): 11 abstract methods, catalog (2), cart (4), preferences (1),\n  orders (2), policies (1), fulfillment (1); `get_account_context`, `get_disclosure`, and `checkout_handoff` have defaults. No method\n  places an order or moves money; `checkout_handoff` names where payment happens (a hosted checkout URL, or one per seller) and the host renders it.\n- `MerchantBackend` (`merchant_agent\u002Fbackend.py`): 16 abstract methods, reads (8: performance 3, catalog 2,\n  inventory and order health 2, pricing 1) and the change lifecycle (8: five `stage_*`, `get_pending_changes`,\n  `apply_change`, `discard_change`); `execute_analysis_query`, `get_analysis_schema`, `get_merchant_context` have defaults.\n- A product or listing sold by size, color, or tier is a family record carrying `options`, its variants each a\n  record with `option_values` and `variant_of`; the cart and price or restock writes take a variant's id and the\n  gates hold a family's. `docs\u002Fbackends.md` has the mapping from common catalog models and what to return for a figure\n  the platform cannot supply.\n- Eligibility, pricing, inventory, and quantity rules are enforced in the backend; the gates check provenance and\n  caps (commerce-trust-safety). Everything a backend returns is fenced before the model reads it.\n- A system that is not wired yet is a method that fails (merchant methods raise `ChangeNotApplicable` from\n  `merchant_agent\u002Fchanges.py`); the tool stays registered, so prompt bytes do not change. A system the business\n  does not have at all is absent: an `enable_*` switch on the role's config turned off, which removes its tools\n  (`absent_tools()`), prompt lines, and grounding rule on all three paths, and the executor refuses those names;\n  the skills that need it are parked under `skills\u002F_staged\u002F`, and the bytes are then fixed for that deployment. `NotOffered` (`shopping_agent\u002Fbackend.py`) is the narrower signal:\n  one item or seller the store does not serve while the system itself is on.\n- `executor_class=` on `ShoppingAgent` \u002F `MerchantAgent`, the SDK toolsets, and the MCP servers' `build_server`\n  takes a subclass of the role's executor for a deployment's own `domain_error` mapping or result wording.\n\n## Delegates\n\nA second model call takes one shape: a `DelegateExtension` (`commerce_common\u002Fdelegation.py`). It receives a brief\nand the handles in `DelegationContext`, never the conversation or the executor; it returns one result validated\nagainst `result_model`; it cannot write, present, or call a delegate; `_run_delegate` in `commerce_common\u002Fexecution.py`\ncaps its calls per turn at `max_delegate_calls_per_turn`. `MerchantAgent` takes `extra_delegates`, and its\nbuilt-in instance is `run_analysis` (commerce-merchant-operations). `ShoppingAgent` registers none.\n\n## Model fields\n\n`model` runs the turn loop and `memory_model` the post-turn extraction (`BaseAgentConfig` in `commerce_common\u002Fconfig.py`);\n`analysis_model` runs the delegate, and `None` means `model` (`MerchantAgentConfig` in `merchant_agent\u002Fconfig.py`). The\nSDK runtimes copy `config.model` into the options; the manifests set `model:`. Choose the values with your own evals.\n\n## Do not\n\n- Vary `tools[]` or the static prompt by request (commerce-prompt-caching).\n- Put a most-turns rule in a skill, or a one-tool rule in the prompt.\n- Enforce a cap or a permission in prompt text alone; it belongs in the executor or the backend.\n- Let the model author a price, a figure, or a term; components are joined from server records (commerce-ui-tools).\n- Add a domain tool to a core package; a vertical adds UI through `PresentationExtension` and keeps the rest in its own code.\n",{"data":38,"body":39},{"name":4,"description":6},{"type":40,"children":41},"root",[42,51,120,127,236,242,363,368,374,509,515,520,784,790,878,884,970,976],{"type":43,"tag":44,"props":45,"children":47},"element","h1",{"id":46},"commerce-agent-architecture",[48],{"type":49,"value":50},"text","Commerce agent architecture",{"type":43,"tag":52,"props":53,"children":54},"p",{},[55,57,64,66,72,74,80,82,88,89,95,96,102,104,110,112,118],{"type":49,"value":56},"Paths are in the reference repo: ",{"type":43,"tag":58,"props":59,"children":61},"code",{"className":60},[],[62],{"type":49,"value":63},"commerce_common\u002F",{"type":49,"value":65}," is ",{"type":43,"tag":58,"props":67,"children":69},{"className":68},[],[70],{"type":49,"value":71},"commerce-common\u002Fcommerce_common\u002F",{"type":49,"value":73},", ",{"type":43,"tag":58,"props":75,"children":77},{"className":76},[],[78],{"type":49,"value":79},"shopping_agent\u002F",{"type":49,"value":81}," is\n",{"type":43,"tag":58,"props":83,"children":85},{"className":84},[],[86],{"type":49,"value":87},"shopping-agent\u002Fcore\u002Fshopping_agent\u002F",{"type":49,"value":73},{"type":43,"tag":58,"props":90,"children":92},{"className":91},[],[93],{"type":49,"value":94},"merchant_agent\u002F",{"type":49,"value":65},{"type":43,"tag":58,"props":97,"children":99},{"className":98},[],[100],{"type":49,"value":101},"merchant-agent\u002Fcore\u002Fmerchant_agent\u002F",{"type":49,"value":103},", and the runtimes are\n",{"type":43,"tag":58,"props":105,"children":107},{"className":106},[],[108],{"type":49,"value":109},"shopping-agent\u002Fruntime-messages-api\u002Fshopping_agent_runtime\u002F",{"type":49,"value":111}," and ",{"type":43,"tag":58,"props":113,"children":115},{"className":114},[],[116],{"type":49,"value":117},"merchant-agent\u002Fruntime-messages-api\u002Fmerchant_agent_runtime\u002F",{"type":49,"value":119},".",{"type":43,"tag":121,"props":122,"children":124},"h2",{"id":123},"one-loop",[125],{"type":49,"value":126},"One loop",{"type":43,"tag":128,"props":129,"children":130},"ul",{},[131,161],{"type":43,"tag":132,"props":133,"children":134},"li",{},[135,137,143,145,151,153,159],{"type":49,"value":136},"One model owns the conversation. A turn is ",{"type":43,"tag":58,"props":138,"children":140},{"className":139},[],[141],{"type":49,"value":142},"ShoppingAgent.stream_turn",{"type":49,"value":144}," or ",{"type":43,"tag":58,"props":146,"children":148},{"className":147},[],[149],{"type":49,"value":150},"MerchantAgent.stream_turn",{"type":49,"value":152}," (each\nruntime's ",{"type":43,"tag":58,"props":154,"children":156},{"className":155},[],[157],{"type":49,"value":158},"orchestrator.py",{"type":49,"value":160},"): the model reads the cached prompt, calls tools (one round's calls run\nconcurrently), and ends with text plus presentation calls. There is no router, classifier, or hand-off.",{"type":43,"tag":132,"props":162,"children":163},{},[164,166,172,174,180,182,188,189,195,197,203,204,210,212,218,220,226,228,234],{"type":49,"value":165},"Every call on every path runs through the role executor (",{"type":43,"tag":58,"props":167,"children":169},{"className":168},[],[170],{"type":49,"value":171},"ShoppingToolExecutor",{"type":49,"value":173}," in ",{"type":43,"tag":58,"props":175,"children":177},{"className":176},[],[178],{"type":49,"value":179},"shopping_agent\u002Fexecutor.py",{"type":49,"value":181},",\n",{"type":43,"tag":58,"props":183,"children":185},{"className":184},[],[186],{"type":49,"value":187},"MerchantToolExecutor",{"type":49,"value":173},{"type":43,"tag":58,"props":190,"children":192},{"className":191},[],[193],{"type":49,"value":194},"merchant_agent\u002Fexecutor.py",{"type":49,"value":196},") over ",{"type":43,"tag":58,"props":198,"children":200},{"className":199},[],[201],{"type":49,"value":202},"BaseToolExecutor",{"type":49,"value":173},{"type":43,"tag":58,"props":205,"children":207},{"className":206},[],[208],{"type":49,"value":209},"commerce_common\u002Fexecution.py",{"type":49,"value":211},".\nThe Messages API runtime, the Agent SDK toolset, and the MCP server call the same ",{"type":43,"tag":58,"props":213,"children":215},{"className":214},[],[216],{"type":49,"value":217},"execute",{"type":49,"value":219},", which never raises;\nafter ",{"type":43,"tag":58,"props":221,"children":223},{"className":222},[],[224],{"type":49,"value":225},"max_tool_iterations",{"type":49,"value":227}," rounds (",{"type":43,"tag":58,"props":229,"children":231},{"className":230},[],[232],{"type":49,"value":233},"commerce_common\u002Fconfig.py",{"type":49,"value":235},") the runtime forces a round without tools.",{"type":43,"tag":121,"props":237,"children":239},{"id":238},"where-a-rule-lives",[240],{"type":49,"value":241},"Where a rule lives",{"type":43,"tag":243,"props":244,"children":245},"table",{},[246,270],{"type":43,"tag":247,"props":248,"children":249},"thead",{},[250],{"type":43,"tag":251,"props":252,"children":253},"tr",{},[254,260,265],{"type":43,"tag":255,"props":256,"children":257},"th",{},[258],{"type":49,"value":259},"A rule that applies",{"type":43,"tag":255,"props":261,"children":262},{},[263],{"type":49,"value":264},"Lives in",{"type":43,"tag":255,"props":266,"children":267},{},[268],{"type":49,"value":269},"Reference",{"type":43,"tag":271,"props":272,"children":273},"tbody",{},[274,305,334],{"type":43,"tag":251,"props":275,"children":276},{},[277,283,288],{"type":43,"tag":278,"props":279,"children":280},"td",{},[281],{"type":49,"value":282},"While one call's arguments are being filled in",{"type":43,"tag":278,"props":284,"children":285},{},[286],{"type":49,"value":287},"That tool's description",{"type":43,"tag":278,"props":289,"children":290},{},[291,297,299],{"type":43,"tag":58,"props":292,"children":294},{"className":293},[],[295],{"type":49,"value":296},"build_tools",{"type":49,"value":298}," in each role's ",{"type":43,"tag":58,"props":300,"children":302},{"className":301},[],[303],{"type":49,"value":304},"tools\u002Fregistry.py",{"type":43,"tag":251,"props":306,"children":307},{},[308,313,318],{"type":43,"tag":278,"props":309,"children":310},{},[311],{"type":49,"value":312},"On most turns: cart and checkout, the staged-write contract, presentation grammar, tool order, trust rules",{"type":43,"tag":278,"props":314,"children":315},{},[316],{"type":49,"value":317},"The static prompt",{"type":43,"tag":278,"props":319,"children":320},{},[321,327,328],{"type":43,"tag":58,"props":322,"children":324},{"className":323},[],[325],{"type":49,"value":326},"build_static_system",{"type":49,"value":298},{"type":43,"tag":58,"props":329,"children":331},{"className":330},[],[332],{"type":49,"value":333},"prompt.py",{"type":43,"tag":251,"props":335,"children":336},{},[337,342,347],{"type":43,"tag":278,"props":338,"children":339},{},[340],{"type":49,"value":341},"On the minority of requests that need a multi-step procedure",{"type":43,"tag":278,"props":343,"children":344},{},[345],{"type":49,"value":346},"A skill, loaded on demand",{"type":43,"tag":278,"props":348,"children":349},{},[350,356,357],{"type":43,"tag":58,"props":351,"children":353},{"className":352},[],[354],{"type":49,"value":355},"shopping-agent\u002Fskills\u002F",{"type":49,"value":73},{"type":43,"tag":58,"props":358,"children":360},{"className":359},[],[361],{"type":49,"value":362},"merchant-agent\u002Fskills\u002F",{"type":43,"tag":52,"props":364,"children":365},{},[366],{"type":49,"value":367},"A rule the core journey needs on most conversations moves a layer down. This table is the one statement of the layering.",{"type":43,"tag":121,"props":369,"children":371},{"id":370},"skills",[372],{"type":49,"value":373},"Skills",{"type":43,"tag":128,"props":375,"children":376},{},[377,421],{"type":43,"tag":132,"props":378,"children":379},{},[380,382,388,390,396,397,403,405,411,413,419],{"type":49,"value":381},"A skill is a directory holding ",{"type":43,"tag":58,"props":383,"children":385},{"className":384},[],[386],{"type":49,"value":387},"SKILL.md",{"type":49,"value":389},": frontmatter ",{"type":43,"tag":58,"props":391,"children":393},{"className":392},[],[394],{"type":49,"value":395},"name",{"type":49,"value":111},{"type":43,"tag":58,"props":398,"children":400},{"className":399},[],[401],{"type":49,"value":402},"description",{"type":49,"value":404},", then the body (",{"type":43,"tag":58,"props":406,"children":408},{"className":407},[],[409],{"type":49,"value":410},"parse_skill_md",{"type":49,"value":412},"\nin ",{"type":43,"tag":58,"props":414,"children":416},{"className":415},[],[417],{"type":49,"value":418},"commerce_common\u002Fskills.py",{"type":49,"value":420},"). The description names the request class; it carries no sample utterances.",{"type":43,"tag":132,"props":422,"children":423},{},[424,426,432,434,440,442,448,449,454,456,462,464,470,471,477,479,485,487,493,495,501,502,508],{"type":49,"value":425},"The static prompt carries the index alone (",{"type":43,"tag":58,"props":427,"children":429},{"className":428},[],[430],{"type":49,"value":431},"SkillRegistry.index_block",{"type":49,"value":433},", sorted by name) and ",{"type":43,"tag":58,"props":435,"children":437},{"className":436},[],[438],{"type":49,"value":439},"load_skill",{"type":49,"value":441}," returns\none body (",{"type":43,"tag":58,"props":443,"children":445},{"className":444},[],[446],{"type":49,"value":447},"_load_skill",{"type":49,"value":173},{"type":43,"tag":58,"props":450,"children":452},{"className":451},[],[453],{"type":49,"value":209},{"type":49,"value":455},"). On the Agent SDK, ",{"type":43,"tag":58,"props":457,"children":459},{"className":458},[],[460],{"type":49,"value":461},"ensure_project_skills",{"type":49,"value":463}," links the\nsame directories into ",{"type":43,"tag":58,"props":465,"children":467},{"className":466},[],[468],{"type":49,"value":469},".claude\u002Fskills\u002F",{"type":49,"value":111},{"type":43,"tag":58,"props":472,"children":474},{"className":473},[],[475],{"type":49,"value":476},"SKILL_TOOL_ADAPTER",{"type":49,"value":478}," points the model at the SDK's ",{"type":43,"tag":58,"props":480,"children":482},{"className":481},[],[483],{"type":49,"value":484},"Skill",{"type":49,"value":486}," tool\n(",{"type":43,"tag":58,"props":488,"children":490},{"className":489},[],[491],{"type":49,"value":492},"commerce_common\u002Fagent_sdk.py",{"type":49,"value":494},"); the hosted manifests list the directories under ",{"type":43,"tag":58,"props":496,"children":498},{"className":497},[],[499],{"type":49,"value":500},"skills[]",{"type":49,"value":173},{"type":43,"tag":58,"props":503,"children":505},{"className":504},[],[506],{"type":49,"value":507},"agent.yaml",{"type":49,"value":119},{"type":43,"tag":121,"props":510,"children":512},{"id":511},"the-backend-interface",[513],{"type":49,"value":514},"The backend interface",{"type":43,"tag":52,"props":516,"children":517},{},[518],{"type":49,"value":519},"A deployment implements one abstract class per role; nothing else reaches its systems.",{"type":43,"tag":128,"props":521,"children":522},{},[523,572,641,677,682,741],{"type":43,"tag":132,"props":524,"children":525},{},[526,532,534,540,542,548,549,555,557,563,565,570],{"type":43,"tag":58,"props":527,"children":529},{"className":528},[],[530],{"type":49,"value":531},"StorefrontBackend",{"type":49,"value":533}," (",{"type":43,"tag":58,"props":535,"children":537},{"className":536},[],[538],{"type":49,"value":539},"shopping_agent\u002Fbackend.py",{"type":49,"value":541},"): 11 abstract methods, catalog (2), cart (4), preferences (1),\norders (2), policies (1), fulfillment (1); ",{"type":43,"tag":58,"props":543,"children":545},{"className":544},[],[546],{"type":49,"value":547},"get_account_context",{"type":49,"value":73},{"type":43,"tag":58,"props":550,"children":552},{"className":551},[],[553],{"type":49,"value":554},"get_disclosure",{"type":49,"value":556},", and ",{"type":43,"tag":58,"props":558,"children":560},{"className":559},[],[561],{"type":49,"value":562},"checkout_handoff",{"type":49,"value":564}," have defaults. No method\nplaces an order or moves money; ",{"type":43,"tag":58,"props":566,"children":568},{"className":567},[],[569],{"type":49,"value":562},{"type":49,"value":571}," names where payment happens (a hosted checkout URL, or one per seller) and the host renders it.",{"type":43,"tag":132,"props":573,"children":574},{},[575,581,582,588,590,596,597,603,604,610,611,617,619,625,626,632,633,639],{"type":43,"tag":58,"props":576,"children":578},{"className":577},[],[579],{"type":49,"value":580},"MerchantBackend",{"type":49,"value":533},{"type":43,"tag":58,"props":583,"children":585},{"className":584},[],[586],{"type":49,"value":587},"merchant_agent\u002Fbackend.py",{"type":49,"value":589},"): 16 abstract methods, reads (8: performance 3, catalog 2,\ninventory and order health 2, pricing 1) and the change lifecycle (8: five ",{"type":43,"tag":58,"props":591,"children":593},{"className":592},[],[594],{"type":49,"value":595},"stage_*",{"type":49,"value":73},{"type":43,"tag":58,"props":598,"children":600},{"className":599},[],[601],{"type":49,"value":602},"get_pending_changes",{"type":49,"value":181},{"type":43,"tag":58,"props":605,"children":607},{"className":606},[],[608],{"type":49,"value":609},"apply_change",{"type":49,"value":73},{"type":43,"tag":58,"props":612,"children":614},{"className":613},[],[615],{"type":49,"value":616},"discard_change",{"type":49,"value":618},"); ",{"type":43,"tag":58,"props":620,"children":622},{"className":621},[],[623],{"type":49,"value":624},"execute_analysis_query",{"type":49,"value":73},{"type":43,"tag":58,"props":627,"children":629},{"className":628},[],[630],{"type":49,"value":631},"get_analysis_schema",{"type":49,"value":73},{"type":43,"tag":58,"props":634,"children":636},{"className":635},[],[637],{"type":49,"value":638},"get_merchant_context",{"type":49,"value":640}," have defaults.",{"type":43,"tag":132,"props":642,"children":643},{},[644,646,652,654,660,661,667,669,675],{"type":49,"value":645},"A product or listing sold by size, color, or tier is a family record carrying ",{"type":43,"tag":58,"props":647,"children":649},{"className":648},[],[650],{"type":49,"value":651},"options",{"type":49,"value":653},", its variants each a\nrecord with ",{"type":43,"tag":58,"props":655,"children":657},{"className":656},[],[658],{"type":49,"value":659},"option_values",{"type":49,"value":111},{"type":43,"tag":58,"props":662,"children":664},{"className":663},[],[665],{"type":49,"value":666},"variant_of",{"type":49,"value":668},"; the cart and price or restock writes take a variant's id and the\ngates hold a family's. ",{"type":43,"tag":58,"props":670,"children":672},{"className":671},[],[673],{"type":49,"value":674},"docs\u002Fbackends.md",{"type":49,"value":676}," has the mapping from common catalog models and what to return for a figure\nthe platform cannot supply.",{"type":43,"tag":132,"props":678,"children":679},{},[680],{"type":49,"value":681},"Eligibility, pricing, inventory, and quantity rules are enforced in the backend; the gates check provenance and\ncaps (commerce-trust-safety). Everything a backend returns is fenced before the model reads it.",{"type":43,"tag":132,"props":683,"children":684},{},[685,687,693,695,701,703,709,711,717,719,725,727,733,734,739],{"type":49,"value":686},"A system that is not wired yet is a method that fails (merchant methods raise ",{"type":43,"tag":58,"props":688,"children":690},{"className":689},[],[691],{"type":49,"value":692},"ChangeNotApplicable",{"type":49,"value":694}," from\n",{"type":43,"tag":58,"props":696,"children":698},{"className":697},[],[699],{"type":49,"value":700},"merchant_agent\u002Fchanges.py",{"type":49,"value":702},"); the tool stays registered, so prompt bytes do not change. A system the business\ndoes not have at all is absent: an ",{"type":43,"tag":58,"props":704,"children":706},{"className":705},[],[707],{"type":49,"value":708},"enable_*",{"type":49,"value":710}," switch on the role's config turned off, which removes its tools\n(",{"type":43,"tag":58,"props":712,"children":714},{"className":713},[],[715],{"type":49,"value":716},"absent_tools()",{"type":49,"value":718},"), prompt lines, and grounding rule on all three paths, and the executor refuses those names;\nthe skills that need it are parked under ",{"type":43,"tag":58,"props":720,"children":722},{"className":721},[],[723],{"type":49,"value":724},"skills\u002F_staged\u002F",{"type":49,"value":726},", and the bytes are then fixed for that deployment. ",{"type":43,"tag":58,"props":728,"children":730},{"className":729},[],[731],{"type":49,"value":732},"NotOffered",{"type":49,"value":533},{"type":43,"tag":58,"props":735,"children":737},{"className":736},[],[738],{"type":49,"value":539},{"type":49,"value":740},") is the narrower signal:\none item or seller the store does not serve while the system itself is on.",{"type":43,"tag":132,"props":742,"children":743},{},[744,750,752,758,760,766,768,774,776,782],{"type":43,"tag":58,"props":745,"children":747},{"className":746},[],[748],{"type":49,"value":749},"executor_class=",{"type":49,"value":751}," on ",{"type":43,"tag":58,"props":753,"children":755},{"className":754},[],[756],{"type":49,"value":757},"ShoppingAgent",{"type":49,"value":759}," \u002F ",{"type":43,"tag":58,"props":761,"children":763},{"className":762},[],[764],{"type":49,"value":765},"MerchantAgent",{"type":49,"value":767},", the SDK toolsets, and the MCP servers' ",{"type":43,"tag":58,"props":769,"children":771},{"className":770},[],[772],{"type":49,"value":773},"build_server",{"type":49,"value":775},"\ntakes a subclass of the role's executor for a deployment's own ",{"type":43,"tag":58,"props":777,"children":779},{"className":778},[],[780],{"type":49,"value":781},"domain_error",{"type":49,"value":783}," mapping or result wording.",{"type":43,"tag":121,"props":785,"children":787},{"id":786},"delegates",[788],{"type":49,"value":789},"Delegates",{"type":43,"tag":52,"props":791,"children":792},{},[793,795,801,802,808,810,816,818,824,826,832,833,838,840,846,848,853,855,861,863,869,871,876],{"type":49,"value":794},"A second model call takes one shape: a ",{"type":43,"tag":58,"props":796,"children":798},{"className":797},[],[799],{"type":49,"value":800},"DelegateExtension",{"type":49,"value":533},{"type":43,"tag":58,"props":803,"children":805},{"className":804},[],[806],{"type":49,"value":807},"commerce_common\u002Fdelegation.py",{"type":49,"value":809},"). It receives a brief\nand the handles in ",{"type":43,"tag":58,"props":811,"children":813},{"className":812},[],[814],{"type":49,"value":815},"DelegationContext",{"type":49,"value":817},", never the conversation or the executor; it returns one result validated\nagainst ",{"type":43,"tag":58,"props":819,"children":821},{"className":820},[],[822],{"type":49,"value":823},"result_model",{"type":49,"value":825},"; it cannot write, present, or call a delegate; ",{"type":43,"tag":58,"props":827,"children":829},{"className":828},[],[830],{"type":49,"value":831},"_run_delegate",{"type":49,"value":173},{"type":43,"tag":58,"props":834,"children":836},{"className":835},[],[837],{"type":49,"value":209},{"type":49,"value":839},"\ncaps its calls per turn at ",{"type":43,"tag":58,"props":841,"children":843},{"className":842},[],[844],{"type":49,"value":845},"max_delegate_calls_per_turn",{"type":49,"value":847},". ",{"type":43,"tag":58,"props":849,"children":851},{"className":850},[],[852],{"type":49,"value":765},{"type":49,"value":854}," takes ",{"type":43,"tag":58,"props":856,"children":858},{"className":857},[],[859],{"type":49,"value":860},"extra_delegates",{"type":49,"value":862},", and its\nbuilt-in instance is ",{"type":43,"tag":58,"props":864,"children":866},{"className":865},[],[867],{"type":49,"value":868},"run_analysis",{"type":49,"value":870}," (commerce-merchant-operations). ",{"type":43,"tag":58,"props":872,"children":874},{"className":873},[],[875],{"type":49,"value":757},{"type":49,"value":877}," registers none.",{"type":43,"tag":121,"props":879,"children":881},{"id":880},"model-fields",[882],{"type":49,"value":883},"Model fields",{"type":43,"tag":52,"props":885,"children":886},{},[887,893,895,901,903,909,910,915,917,923,925,931,933,938,939,945,946,952,954,960,962,968],{"type":43,"tag":58,"props":888,"children":890},{"className":889},[],[891],{"type":49,"value":892},"model",{"type":49,"value":894}," runs the turn loop and ",{"type":43,"tag":58,"props":896,"children":898},{"className":897},[],[899],{"type":49,"value":900},"memory_model",{"type":49,"value":902}," the post-turn extraction (",{"type":43,"tag":58,"props":904,"children":906},{"className":905},[],[907],{"type":49,"value":908},"BaseAgentConfig",{"type":49,"value":173},{"type":43,"tag":58,"props":911,"children":913},{"className":912},[],[914],{"type":49,"value":233},{"type":49,"value":916},");\n",{"type":43,"tag":58,"props":918,"children":920},{"className":919},[],[921],{"type":49,"value":922},"analysis_model",{"type":49,"value":924}," runs the delegate, and ",{"type":43,"tag":58,"props":926,"children":928},{"className":927},[],[929],{"type":49,"value":930},"None",{"type":49,"value":932}," means ",{"type":43,"tag":58,"props":934,"children":936},{"className":935},[],[937],{"type":49,"value":892},{"type":49,"value":533},{"type":43,"tag":58,"props":940,"children":942},{"className":941},[],[943],{"type":49,"value":944},"MerchantAgentConfig",{"type":49,"value":173},{"type":43,"tag":58,"props":947,"children":949},{"className":948},[],[950],{"type":49,"value":951},"merchant_agent\u002Fconfig.py",{"type":49,"value":953},"). The\nSDK runtimes copy ",{"type":43,"tag":58,"props":955,"children":957},{"className":956},[],[958],{"type":49,"value":959},"config.model",{"type":49,"value":961}," into the options; the manifests set ",{"type":43,"tag":58,"props":963,"children":965},{"className":964},[],[966],{"type":49,"value":967},"model:",{"type":49,"value":969},". Choose the values with your own evals.",{"type":43,"tag":121,"props":971,"children":973},{"id":972},"do-not",[974],{"type":49,"value":975},"Do not",{"type":43,"tag":128,"props":977,"children":978},{},[979,992,997,1002,1007],{"type":43,"tag":132,"props":980,"children":981},{},[982,984,990],{"type":49,"value":983},"Vary ",{"type":43,"tag":58,"props":985,"children":987},{"className":986},[],[988],{"type":49,"value":989},"tools[]",{"type":49,"value":991}," or the static prompt by request (commerce-prompt-caching).",{"type":43,"tag":132,"props":993,"children":994},{},[995],{"type":49,"value":996},"Put a most-turns rule in a skill, or a one-tool rule in the prompt.",{"type":43,"tag":132,"props":998,"children":999},{},[1000],{"type":49,"value":1001},"Enforce a cap or a permission in prompt text alone; it belongs in the executor or the backend.",{"type":43,"tag":132,"props":1003,"children":1004},{},[1005],{"type":49,"value":1006},"Let the model author a price, a figure, or a term; components are joined from server records (commerce-ui-tools).",{"type":43,"tag":132,"props":1008,"children":1009},{},[1010,1012,1018],{"type":49,"value":1011},"Add a domain tool to a core package; a vertical adds UI through ",{"type":43,"tag":58,"props":1013,"children":1015},{"className":1014},[],[1016],{"type":49,"value":1017},"PresentationExtension",{"type":49,"value":1019}," and keeps the rest in its own code.",{"items":1021,"total":1205},[1022,1038,1057,1071,1083,1100,1116,1127,1148,1168,1182,1197],{"slug":1023,"name":1023,"fn":1024,"description":1025,"org":1026,"tags":1027,"stars":1035,"repoUrl":1036,"updatedAt":1037},"academy-guide","recommend Claude Academy resources","Stop and check this skill before finishing any reply to a question about how to use Claude or a Claude product — it recommends matching courses, tutorials, and use cases from Claude Academy (academy.claude.com), Anthropic's learning hub. Trigger on: \"how do I\", \"how can I\", \"getting started with\", \"what can Claude do\", \"teach me\", \"learn to use\"; questions about artifacts, projects, skills, plugins, connectors, MCP; requests about rolling Claude out to a team, class, or organization; and any ask for training materials, onboarding content, or learning resources. Use it when the user is learning how to use a feature or product — not when they are mid-task and just want the task done. This skill composes with other skills: after consulting product documentation to answer how a Claude feature works, also check here for a matching course or tutorial — a docs-grounded answer and an Academy recommendation belong together. Only recommend on a strong match; never invent Academy content.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1028,1029,1032],{"name":9,"slug":8,"type":16},{"name":1030,"slug":1031,"type":16},"Documentation","documentation",{"name":1033,"slug":1034,"type":16},"Education","education",161831,"https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fskills","2026-08-19T03:59:01.021254",{"slug":1039,"name":1039,"fn":1040,"description":1041,"org":1042,"tags":1043,"stars":1035,"repoUrl":1036,"updatedAt":1056},"algorithmic-art","create algorithmic art with p5.js","Creating algorithmic art using p5.js with seeded randomness and interactive parameter exploration. Use this when users request creating art using code, generative art, algorithmic art, flow fields, or particle systems. Create original algorithmic art rather than copying existing artists' work to avoid copyright violations.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1044,1047,1050,1053],{"name":1045,"slug":1046,"type":16},"Creative","creative",{"name":1048,"slug":1049,"type":16},"Design","design",{"name":1051,"slug":1052,"type":16},"Generative Art","generative-art",{"name":1054,"slug":1055,"type":16},"JavaScript","javascript","2026-04-06T17:56:15.455818",{"slug":1058,"name":1058,"fn":1059,"description":1060,"org":1061,"tags":1062,"stars":1035,"repoUrl":1036,"updatedAt":1070},"brand-guidelines","apply Anthropic brand colors and typography","Applies Anthropic's official brand colors and typography to any sort of artifact that may benefit from having Anthropic's look-and-feel. Use it when brand colors or style guidelines, visual formatting, or company design standards apply.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1063,1066,1067],{"name":1064,"slug":1065,"type":16},"Branding","branding",{"name":1048,"slug":1049,"type":16},{"name":1068,"slug":1069,"type":16},"Typography","typography","2026-04-06T17:56:05.042852",{"slug":1072,"name":1072,"fn":1073,"description":1074,"org":1075,"tags":1076,"stars":1035,"repoUrl":1036,"updatedAt":1082},"canvas-design","create posters and visual art as PNG or PDF","Create beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1077,1078,1079],{"name":1045,"slug":1046,"type":16},{"name":1048,"slug":1049,"type":16},{"name":1080,"slug":1081,"type":16},"PDF","pdf","2026-04-06T17:56:03.794732",{"slug":1084,"name":1084,"fn":1085,"description":1086,"org":1087,"tags":1088,"stars":1035,"repoUrl":1036,"updatedAt":1099},"claude-api","build apps with the Claude API","Reference for the Claude API \u002F Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration.\nTRIGGER — read BEFORE opening the target file; don't skip because it \"looks like a one-liner\" — whenever: the prompt names Claude\u002FAnthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing\u002Fmodel choice\u002Flimits\u002Fcaching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent\u002FMCP\u002Ftool-definition\u002Fmulti-agent\u002FRAG\u002FLLM-judge\u002Fcomputer-use; generate\u002Fsummarize\u002Fextract\u002Fclassify\u002Frewrite\u002Fconverse over NL; debugging refusals\u002Fcutoffs\u002Fstreaming\u002Ftool-calls\u002Ftokens).\nSKIP only when another provider is being worked on (overrides all triggers): OpenAI\u002FGPT\u002FGemini\u002FLlama\u002FMistral\u002FCohere\u002FOllama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file).",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1089,1090,1091,1094,1096],{"name":21,"slug":22,"type":16},{"name":9,"slug":8,"type":16},{"name":1092,"slug":1093,"type":16},"Anthropic SDK","anthropic-sdk",{"name":1095,"slug":1084,"type":16},"Claude API",{"name":1097,"slug":1098,"type":16},"LLM","llm","2026-09-04T07:28:24.898544",{"slug":1101,"name":1101,"fn":1102,"description":1103,"org":1104,"tags":1105,"stars":1035,"repoUrl":1036,"updatedAt":1115},"discernment-nudge","provide discernment nudges for user decisions","After you give a substantive answer or draft that the user may act on — advice or recommendations, drafted artifacts such as goals, plans, pitches, proposals, or emails, estimates or projections, analysis or interpretation of data, factual claims they may rely on, or a multi-step argument — invoke this skill BEFORE finalizing your reply and then, if it applies, append 2-3 short follow-up questions, each tied to something specific in what you just produced, that help the user check key facts, probe the reasoning or assumptions, and notice missing context. Do this at most once per conversation. Skip it when the user asked a trivial how-to or simple lookup, wants a purely educational explanation, asked you only to format, convert, or assemble a file from content they provided, is writing code they will run, is doing creative writing or casual chat, or already asked you to double-check, cite, or review — the skill file explains these boundaries and the exact output format.\n",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1106,1109,1112],{"name":1107,"slug":1108,"type":16},"Coaching","coaching",{"name":1110,"slug":1111,"type":16},"Productivity","productivity",{"name":1113,"slug":1114,"type":16},"Strategy","strategy","2026-08-19T03:59:01.539281",{"slug":1117,"name":1117,"fn":1118,"description":1119,"org":1120,"tags":1121,"stars":1035,"repoUrl":1036,"updatedAt":1126},"doc-coauthoring","co-author documentation and technical specs","Guide users through a structured workflow for co-authoring documentation. Use when user wants to write documentation, proposals, technical specs, decision docs, or similar structured content. This workflow helps users efficiently transfer context, refine content through iteration, and verify the doc works for readers. Trigger when user mentions writing docs, creating proposals, drafting specs, or similar documentation tasks.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1122,1123],{"name":1030,"slug":1031,"type":16},{"name":1124,"slug":1125,"type":16},"Technical Writing","technical-writing","2026-04-06T17:56:14.18897",{"slug":1128,"name":1128,"fn":1129,"description":1130,"org":1131,"tags":1132,"stars":1035,"repoUrl":1036,"updatedAt":1147},"docx","create and edit Word documents","Use this skill whenever the user wants to create, read, edit, or manipulate Word documents (.docx files) or Word templates (.dotx files). Triggers include: any mention of 'Word doc', 'word document', '.docx', '.dotx', or requests to produce professional documents with formatting like tables of contents, headings, page numbers, or letterheads. Also use when extracting or reorganizing content from .docx or .dotx files, inserting or replacing images in documents, performing find-and-replace in Word files, working with tracked changes or comments, or converting content into a polished Word document. If the user asks for a 'report', 'memo', 'letter', 'template', or similar deliverable as a Word or .docx file, use this skill. Do NOT use for PDFs, spreadsheets, Google Docs, or general coding tasks unrelated to document generation.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1133,1136,1138,1141,1144],{"name":1134,"slug":1135,"type":16},"Documents","documents",{"name":1137,"slug":1128,"type":16},"DOCX",{"name":1139,"slug":1140,"type":16},"Office","office",{"name":1142,"slug":1143,"type":16},"Templates","templates",{"name":1145,"slug":1146,"type":16},"Word","word","2026-07-18T05:16:23.136271",{"slug":1149,"name":1149,"fn":1150,"description":1151,"org":1152,"tags":1153,"stars":1035,"repoUrl":1036,"updatedAt":1167},"frontend-design","design production-grade frontend interfaces","Guidance for distinctive, intentional visual design when building new UI or reshaping an existing one. Helps with aesthetic direction, typography, and making choices that don't read as templated defaults.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1154,1155,1158,1161,1164],{"name":1048,"slug":1049,"type":16},{"name":1156,"slug":1157,"type":16},"Frontend","frontend",{"name":1159,"slug":1160,"type":16},"React","react",{"name":1162,"slug":1163,"type":16},"Tailwind CSS","tailwind-css",{"name":1165,"slug":1166,"type":16},"UI Components","ui-components","2026-09-04T07:28:23.795756",{"slug":1169,"name":1169,"fn":1170,"description":1171,"org":1172,"tags":1173,"stars":1035,"repoUrl":1036,"updatedAt":1181},"internal-comms","write internal company communications","A set of resources to help me write all kinds of internal communications, using the formats that my company likes to use. Claude should use this skill whenever asked to write some sort of internal communications (status reports, leadership updates, 3P updates, company newsletters, FAQs, incident reports, project updates, etc.).",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1174,1177,1178],{"name":1175,"slug":1176,"type":16},"Communications","communications",{"name":1142,"slug":1143,"type":16},{"name":1179,"slug":1180,"type":16},"Writing","writing","2026-04-06T17:56:20.695522",{"slug":1183,"name":1183,"fn":1184,"description":1185,"org":1186,"tags":1187,"stars":1035,"repoUrl":1036,"updatedAt":1196},"mcp-builder","build MCP servers","Guide for creating high-quality MCP (Model Context Protocol) servers that enable LLMs to interact with external services through well-designed tools. Use when building MCP servers to integrate external APIs or services, whether in Python (FastMCP) or Node\u002FTypeScript (MCP SDK).",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1188,1189,1192,1193],{"name":21,"slug":22,"type":16},{"name":1190,"slug":1191,"type":16},"API Development","api-development",{"name":1097,"slug":1098,"type":16},{"name":1194,"slug":1195,"type":16},"MCP","mcp","2026-04-06T17:56:10.357665",{"slug":1081,"name":1081,"fn":1198,"description":1199,"org":1200,"tags":1201,"stars":1035,"repoUrl":1036,"updatedAt":1204},"read edit and manipulate PDF files","Use this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text\u002Ftables from PDFs, combining or merging multiple PDFs into one, splitting PDFs apart, rotating pages, adding watermarks, creating new PDFs, filling PDF forms, encrypting\u002Fdecrypting PDFs, extracting images, and OCR on scanned PDFs to make them searchable. If the user mentions a .pdf file or asks to produce one, use this skill.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1202,1203],{"name":1134,"slug":1135,"type":16},{"name":1080,"slug":1081,"type":16},"2026-04-06T17:56:02.483316",521,{"items":1207,"total":1305},[1208,1225,1232,1247,1262,1277,1294],{"slug":1209,"name":1209,"fn":1210,"description":1211,"org":1212,"tags":1213,"stars":26,"repoUrl":27,"updatedAt":1224},"catalog-listings","create and improve e-commerce listing content","Creating and improving listing content, covering titles, descriptions, attribute completeness, categorization fixes, image callouts written as text, edits written from material the operator supplies, and content-quality audits and bulk fixes across many listings. Not needed for questions about how a listing is performing.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1214,1217,1220,1221],{"name":1215,"slug":1216,"type":16},"Content Creation","content-creation",{"name":1218,"slug":1219,"type":16},"Content Strategy","content-strategy",{"name":18,"slug":19,"type":16},{"name":1222,"slug":1223,"type":16},"Marketing","marketing","2026-09-04T08:00:35.843558",{"slug":4,"name":4,"fn":5,"description":6,"org":1226,"tags":1227,"stars":26,"repoUrl":27,"updatedAt":28},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1228,1229,1230,1231],{"name":21,"slug":22,"type":16},{"name":14,"slug":15,"type":16},{"name":18,"slug":19,"type":16},{"name":24,"slug":25,"type":16},{"slug":1233,"name":1233,"fn":1234,"description":1235,"org":1236,"tags":1237,"stars":26,"repoUrl":27,"updatedAt":1246},"commerce-evals","run behavioral evals for commerce agents","Authoring and running behavioral evals for a shopping or merchant agent, covering the case shape, authoring rules, code graders and judges, the run pattern, and poisoned fixtures. Load when writing eval cases or rubrics or deciding how a suite runs.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1238,1239,1240,1243],{"name":21,"slug":22,"type":16},{"name":18,"slug":19,"type":16},{"name":1241,"slug":1242,"type":16},"Evals","evals",{"name":1244,"slug":1245,"type":16},"Testing","testing","2026-09-04T08:00:36.204114",{"slug":1248,"name":1248,"fn":1249,"description":1250,"org":1251,"tags":1252,"stars":26,"repoUrl":27,"updatedAt":1261},"commerce-merchant-operations","manage merchant operations and store rules","The reference merchant agent, covering its flows, staged changes and host approval, metrics grounding, the analysis delegate, store memory, components, and marketplace rules. Load when building or reviewing an agent for the operators of a store, property, subscriber base, or venue.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1253,1254,1255,1258],{"name":21,"slug":22,"type":16},{"name":18,"slug":19,"type":16},{"name":1256,"slug":1257,"type":16},"Management","management",{"name":1259,"slug":1260,"type":16},"Operations","operations","2026-09-04T08:00:35.464555",{"slug":1263,"name":1263,"fn":1264,"description":1265,"org":1266,"tags":1267,"stars":26,"repoUrl":27,"updatedAt":1276},"commerce-prompt-caching","configure commerce agent prompt caching","The reference agents' cache-stable request assembly, covering the static system and per-request context split, the fixed tool list, the rolling conversation breakpoint, which config fields are prompt bytes, and verification. Load when writing or reviewing a commerce agent's prompt assembly, when cache reads are zero, or when a turn's latency or cost is the question.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1268,1269,1270,1273],{"name":21,"slug":22,"type":16},{"name":18,"slug":19,"type":16},{"name":1271,"slug":1272,"type":16},"Performance","performance",{"name":1274,"slug":1275,"type":16},"Prompt Engineering","prompt-engineering","2026-09-04T08:00:39.50364",{"slug":1278,"name":1278,"fn":1279,"description":1280,"org":1281,"tags":1282,"stars":26,"repoUrl":27,"updatedAt":1293},"commerce-trust-safety","enforce trust and safety rules","The rules the reference agents enforce in code for third-party content, writes, grounding, identity, and memory, each with its module, plus two adversarial-eval rules. Load when handling untrusted tool results, guarding a write tool, or scoping what an agent remembers.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1283,1284,1287,1290],{"name":21,"slug":22,"type":16},{"name":1285,"slug":1286,"type":16},"Compliance","compliance",{"name":1288,"slug":1289,"type":16},"Memory","memory",{"name":1291,"slug":1292,"type":16},"Security","security","2026-09-04T08:00:42.609563",{"slug":1295,"name":1295,"fn":1296,"description":1297,"org":1298,"tags":1299,"stars":26,"repoUrl":27,"updatedAt":1304},"commerce-ui-tools","implement commerce agent UI components","The reference presentation-tool contract, covering server-side enrichment, suggestion chips, the event stream, progressive rendering, both roles' built-in components, and adding a vertical component. Load when building or reviewing how a commerce agent's output reaches a frontend.",{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[1300,1301,1302,1303],{"name":1048,"slug":1049,"type":16},{"name":18,"slug":19,"type":16},{"name":1156,"slug":1157,"type":16},{"name":1165,"slug":1166,"type":16},"2026-09-04T08:00:38.801247",16]