[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill-anthropic-commerce-evals":3,"mdc-80ded7-key":37,"related-org-anthropic-commerce-evals":1783,"related-repo-anthropic-commerce-evals":1969},{"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-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},"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},"Evals","evals","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},"Testing","testing",1735,"https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fcommerce-agents","2026-09-04T08:00:36.204114",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-evals","---\nname: commerce-evals\ndescription: 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.\n---\n\n# Commerce agent evals\n\nThe repo ships no eval harness; the suite is yours, because a case only means something against your catalog,\norders, and fixtures. Paths below are in the reference repo: `commerce_common\u002F` is `commerce-common\u002Fcommerce_common\u002F`.\nGate behavior that needs no model (provenance, caps, guardrails, approval) is unit-tested with `FakeClient` in\n`commerce_common\u002Ftesting.py`; evals cover what the model decides.\n\n## The case shape\n\nThis block is the one home of the case shape and the scorer names; `\u002Fauthor-commerce-evals` refers to it.\n\n```json\n{\n  \"id\": \"\u003Cflow>-\u003Cnnn>-\u003Cbehavior>\",\n  \"priority\": \"critical | high | medium | low\",   \"difficulty\": \"easy | medium | hard\",   \"tags\": [\"...\"],\n  \"skip\": \"\u003Creason, when a case cannot run yet>\",\n  \"state\": {\"seen_products\": [\"...\"], \"cart\": [...], \"memory\": [...], \"staged_changes\": [...]},\n  \"turns\": [\"\u003Cthe customer's or operator's message>\", \"...\"],\n  \"expected\": {\n    \"calls_tool\": [\"...\"],            \"calls_one_of\": [\"...\"],          \"never_calls\": [\"...\"],\n    \"first_tool\": \"...\",              \"first_tool_not\": \"...\",\n    \"ui_components\": [\"...\"],         \"no_ui\": true,\n    \"cart_contains\": [\"...\"],         \"cart_item_count\": 0,             \"cart_not_contains\": [\"...\"],\n    \"staged_change_kinds\": [\"...\"],   \"no_applied_changes\": true,\n    \"memory_contains\": [\"...\"],       \"memory_not_contains\": [\"...\"],\n    \"skill_loaded\": \"...\",            \"skill_not_loaded\": \"...\",        \"no_skill_load\": true,\n    \"reply_includes\": [\"...\"],        \"reply_omits\": [\"...\"],           \"max_tool_calls\": 0,\n    \"rubric\": \"PASS if \u003Ccondition>. FAIL if \u003Ccondition>.\"\n  },\n  \"notes\": \"\u003Cwhat the case pins and the fixture fact that decides it>\"\n}\n```\n\n`state` is the precondition; `turns` is one message unless the behavior under test is carrying state across turns;\n`expected` holds only the keys the case is about. Ids in `state` and `expected` are real ids from your fixtures.\n`priority` and `difficulty` let a report say which failures matter; a case that cannot run yet carries `skip`\nwith its reason rather than being deleted.\n\n## Authoring rules\n\n- Preconditions go in injected state (the products already seen, the cart, the memory facts, the staged queue), which\n  the runner loads into the session state and the memory store before the turn. Earlier turns are for state the\n  behavior itself carries, and for nothing else.\n- Every positive has a negative: for each case that asserts a component, a skill load, a gate, a disclosure, or a\n  memory write, a case in the same niche asserts its absence. A refusal case has a should-serve counterpart.\n- Memory is three cases: a remark worth keeping is written; an identifier or excluded content is refused (nothing\n  stored, no error to the person); a stored fact changes the next session's pick.\n- Grade the final tool arguments and the state they produced: the ids on the last presentation call, the fields on\n  the staged change, the cart and the memory store after the turn. The reply's wording is graded only for strings\n  that must or must not appear. Which route the agent took (`skill_loaded`, one named component, `never_calls` on a\n  presentation tool) is asserted only where the route is the behavior (a grounding read first, a write that must\n  never happen); elsewhere `calls_one_of` names the acceptable set. When a live run takes a route the case did not\n  expect and the answer was right, widen the case to the acceptable set; do not re-pin it to the route observed.\n- A rubric is one PASS condition and one FAIL condition that no response satisfies both of; it names the fixture\n  fact that decides it (the updated delivery date, the price today); variants you accept are written into it; it\n  says nothing about tone, length, or the order components appear in.\n- A turn that mentions health, a one-off errand, or hostile content asserts the memory end-state\n  (`memory_not_contains`, or `never_calls` on `save_memory`). A case where a stored fact should change the pick uses a\n  query whose results contain both the item the fact favors and the one it rules out; run the search before writing\n  the case.\n- `max_tool_calls` is set from what a well-behaved agent needs; a multi-item request fans out several searches in one round, and the `present_suggestions` call that ends the turn is not counted.\n\n## Scorers\n\n- Code graders read the events a turn yields (`commerce_common\u002Fstreaming.py`): `tool_call` names and arguments,\n  `tool_result` with `status` `blocked` and the gate in `reason`, `ui` component names, the last `cart_update` or\n  `change_update`, and the reply text. Every key above except `rubric` is a code grader.\n- `rubric` goes to a judge. One judge call per dimension (budget respected, no invented availability, trade-off stated),\n  returning structured output with a verdict and a reason; the transcript is passed to it as quoted material, tool\n  results and component payloads included; when the transcript exceeds the judge's window, truncate from the start so\n  the graded turn survives, and record the truncation on the outcome. Pin the judge model at temperature zero; a change to the judge model or a rubric\n  invalidates every stored verdict scored with it, so the recording carries a fingerprint of both.\n- A judge reply that does not parse into a verdict is a judge failure on the case, kept apart from an agent failure.\n\n## Run pattern\n\n| When | What runs | What decides |\n|---|---|---|\n| Every merge to the agent, a skill, a tool description, or a fixture | The regression set | Each case over several trials; a pass threshold per set |\n| While changing one flow | That flow's targeted set | The failure set, read beside the previous run's as the baseline |\n| Choosing or upgrading a model (commerce-architecture's model fields) | Everything | The two failure sets side by side |\n| In production | A judged sample of live traffic against the same rubrics | Trend per dimension |\n\nDiff failure sets; a topline moving a point between live runs is noise. A case that fails after a change means the\nchange broke the behavior or the case encoded a stale one; fix whichever it is and say which in the commit.\n\n## Poisoned fixtures\n\nListings, reviews, and messages carrying instructions live in eval-only fixtures the runner merges into the backend\nfor the run, under a third-party brand or seller; none of their ids appears in demo data, seeds, or captures. Each\nsuch case asserts the negative in code (`never_calls`, `cart_not_contains`, `no_applied_changes`,\n`memory_not_contains`, `reply_omits`), and every vector it asserts is one the driven turn actually puts in front of\nthe model (a review is only read on a details call). Cover at least an instruction to write to the cart or stage a\nchange, one to remember something, and a false claim (a code, a guarantee). The should-serve counterpart is a\nseparate benign eval-only listing in the same niche, so an agent that refuses everything fails it.\n",{"data":38,"body":39},{"name":4,"description":6},{"type":40,"children":41},"root",[42,51,90,97,110,1351,1409,1415,1511,1517,1619,1625,1730,1735,1741,1777],{"type":43,"tag":44,"props":45,"children":47},"element","h1",{"id":46},"commerce-agent-evals",[48],{"type":49,"value":50},"text","Commerce agent evals",{"type":43,"tag":52,"props":53,"children":54},"p",{},[55,57,64,66,72,74,80,82,88],{"type":49,"value":56},"The repo ships no eval harness; the suite is yours, because a case only means something against your catalog,\norders, and fixtures. Paths below 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},".\nGate behavior that needs no model (provenance, caps, guardrails, approval) is unit-tested with ",{"type":43,"tag":58,"props":75,"children":77},{"className":76},[],[78],{"type":49,"value":79},"FakeClient",{"type":49,"value":81}," in\n",{"type":43,"tag":58,"props":83,"children":85},{"className":84},[],[86],{"type":49,"value":87},"commerce_common\u002Ftesting.py",{"type":49,"value":89},"; evals cover what the model decides.",{"type":43,"tag":91,"props":92,"children":94},"h2",{"id":93},"the-case-shape",[95],{"type":49,"value":96},"The case shape",{"type":43,"tag":52,"props":98,"children":99},{},[100,102,108],{"type":49,"value":101},"This block is the one home of the case shape and the scorer names; ",{"type":43,"tag":58,"props":103,"children":105},{"className":104},[],[106],{"type":49,"value":107},"\u002Fauthor-commerce-evals",{"type":49,"value":109}," refers to it.",{"type":43,"tag":111,"props":112,"children":117},"pre",{"className":113,"code":114,"language":115,"meta":116,"style":116},"language-json shiki shiki-themes material-theme-lighter material-theme material-theme-palenight","{\n  \"id\": \"\u003Cflow>-\u003Cnnn>-\u003Cbehavior>\",\n  \"priority\": \"critical | high | medium | low\",   \"difficulty\": \"easy | medium | hard\",   \"tags\": [\"...\"],\n  \"skip\": \"\u003Creason, when a case cannot run yet>\",\n  \"state\": {\"seen_products\": [\"...\"], \"cart\": [...], \"memory\": [...], \"staged_changes\": [...]},\n  \"turns\": [\"\u003Cthe customer's or operator's message>\", \"...\"],\n  \"expected\": {\n    \"calls_tool\": [\"...\"],            \"calls_one_of\": [\"...\"],          \"never_calls\": [\"...\"],\n    \"first_tool\": \"...\",              \"first_tool_not\": \"...\",\n    \"ui_components\": [\"...\"],         \"no_ui\": true,\n    \"cart_contains\": [\"...\"],         \"cart_item_count\": 0,             \"cart_not_contains\": [\"...\"],\n    \"staged_change_kinds\": [\"...\"],   \"no_applied_changes\": true,\n    \"memory_contains\": [\"...\"],       \"memory_not_contains\": [\"...\"],\n    \"skill_loaded\": \"...\",            \"skill_not_loaded\": \"...\",        \"no_skill_load\": true,\n    \"reply_includes\": [\"...\"],        \"reply_omits\": [\"...\"],           \"max_tool_calls\": 0,\n    \"rubric\": \"PASS if \u003Ccondition>. FAIL if \u003Ccondition>.\"\n  },\n  \"notes\": \"\u003Cwhat the case pins and the fixture fact that decides it>\"\n}\n","json","",[118],{"type":43,"tag":58,"props":119,"children":120},{"__ignoreMap":116},[121,133,178,292,330,484,542,568,686,757,821,927,989,1068,1160,1264,1299,1308,1342],{"type":43,"tag":122,"props":123,"children":126},"span",{"class":124,"line":125},"line",1,[127],{"type":43,"tag":122,"props":128,"children":130},{"style":129},"--shiki-light:#39ADB5;--shiki-default:#89DDFF;--shiki-dark:#89DDFF",[131],{"type":49,"value":132},"{\n",{"type":43,"tag":122,"props":134,"children":136},{"class":124,"line":135},2,[137,142,148,153,158,163,169,173],{"type":43,"tag":122,"props":138,"children":139},{"style":129},[140],{"type":49,"value":141},"  \"",{"type":43,"tag":122,"props":143,"children":145},{"style":144},"--shiki-light:#9C3EDA;--shiki-default:#C792EA;--shiki-dark:#C792EA",[146],{"type":49,"value":147},"id",{"type":43,"tag":122,"props":149,"children":150},{"style":129},[151],{"type":49,"value":152},"\"",{"type":43,"tag":122,"props":154,"children":155},{"style":129},[156],{"type":49,"value":157},":",{"type":43,"tag":122,"props":159,"children":160},{"style":129},[161],{"type":49,"value":162}," \"",{"type":43,"tag":122,"props":164,"children":166},{"style":165},"--shiki-light:#91B859;--shiki-default:#C3E88D;--shiki-dark:#C3E88D",[167],{"type":49,"value":168},"\u003Cflow>-\u003Cnnn>-\u003Cbehavior>",{"type":43,"tag":122,"props":170,"children":171},{"style":129},[172],{"type":49,"value":152},{"type":43,"tag":122,"props":174,"children":175},{"style":129},[176],{"type":49,"value":177},",\n",{"type":43,"tag":122,"props":179,"children":181},{"class":124,"line":180},3,[182,186,191,195,199,203,208,212,217,222,227,231,235,239,244,248,252,256,261,265,269,274,278,283,287],{"type":43,"tag":122,"props":183,"children":184},{"style":129},[185],{"type":49,"value":141},{"type":43,"tag":122,"props":187,"children":188},{"style":144},[189],{"type":49,"value":190},"priority",{"type":43,"tag":122,"props":192,"children":193},{"style":129},[194],{"type":49,"value":152},{"type":43,"tag":122,"props":196,"children":197},{"style":129},[198],{"type":49,"value":157},{"type":43,"tag":122,"props":200,"children":201},{"style":129},[202],{"type":49,"value":162},{"type":43,"tag":122,"props":204,"children":205},{"style":165},[206],{"type":49,"value":207},"critical | high | medium | low",{"type":43,"tag":122,"props":209,"children":210},{"style":129},[211],{"type":49,"value":152},{"type":43,"tag":122,"props":213,"children":214},{"style":129},[215],{"type":49,"value":216},",",{"type":43,"tag":122,"props":218,"children":219},{"style":129},[220],{"type":49,"value":221},"   \"",{"type":43,"tag":122,"props":223,"children":224},{"style":144},[225],{"type":49,"value":226},"difficulty",{"type":43,"tag":122,"props":228,"children":229},{"style":129},[230],{"type":49,"value":152},{"type":43,"tag":122,"props":232,"children":233},{"style":129},[234],{"type":49,"value":157},{"type":43,"tag":122,"props":236,"children":237},{"style":129},[238],{"type":49,"value":162},{"type":43,"tag":122,"props":240,"children":241},{"style":165},[242],{"type":49,"value":243},"easy | medium | hard",{"type":43,"tag":122,"props":245,"children":246},{"style":129},[247],{"type":49,"value":152},{"type":43,"tag":122,"props":249,"children":250},{"style":129},[251],{"type":49,"value":216},{"type":43,"tag":122,"props":253,"children":254},{"style":129},[255],{"type":49,"value":221},{"type":43,"tag":122,"props":257,"children":258},{"style":144},[259],{"type":49,"value":260},"tags",{"type":43,"tag":122,"props":262,"children":263},{"style":129},[264],{"type":49,"value":152},{"type":43,"tag":122,"props":266,"children":267},{"style":129},[268],{"type":49,"value":157},{"type":43,"tag":122,"props":270,"children":271},{"style":129},[272],{"type":49,"value":273}," [",{"type":43,"tag":122,"props":275,"children":276},{"style":129},[277],{"type":49,"value":152},{"type":43,"tag":122,"props":279,"children":280},{"style":165},[281],{"type":49,"value":282},"...",{"type":43,"tag":122,"props":284,"children":285},{"style":129},[286],{"type":49,"value":152},{"type":43,"tag":122,"props":288,"children":289},{"style":129},[290],{"type":49,"value":291},"],\n",{"type":43,"tag":122,"props":293,"children":295},{"class":124,"line":294},4,[296,300,305,309,313,317,322,326],{"type":43,"tag":122,"props":297,"children":298},{"style":129},[299],{"type":49,"value":141},{"type":43,"tag":122,"props":301,"children":302},{"style":144},[303],{"type":49,"value":304},"skip",{"type":43,"tag":122,"props":306,"children":307},{"style":129},[308],{"type":49,"value":152},{"type":43,"tag":122,"props":310,"children":311},{"style":129},[312],{"type":49,"value":157},{"type":43,"tag":122,"props":314,"children":315},{"style":129},[316],{"type":49,"value":162},{"type":43,"tag":122,"props":318,"children":319},{"style":165},[320],{"type":49,"value":321},"\u003Creason, when a case cannot run yet>",{"type":43,"tag":122,"props":323,"children":324},{"style":129},[325],{"type":49,"value":152},{"type":43,"tag":122,"props":327,"children":328},{"style":129},[329],{"type":49,"value":177},{"type":43,"tag":122,"props":331,"children":333},{"class":124,"line":332},5,[334,338,343,347,351,356,360,366,370,374,378,382,386,390,395,399,404,408,412,416,421,425,429,434,438,442,446,450,454,458,463,467,471,475,479],{"type":43,"tag":122,"props":335,"children":336},{"style":129},[337],{"type":49,"value":141},{"type":43,"tag":122,"props":339,"children":340},{"style":144},[341],{"type":49,"value":342},"state",{"type":43,"tag":122,"props":344,"children":345},{"style":129},[346],{"type":49,"value":152},{"type":43,"tag":122,"props":348,"children":349},{"style":129},[350],{"type":49,"value":157},{"type":43,"tag":122,"props":352,"children":353},{"style":129},[354],{"type":49,"value":355}," {",{"type":43,"tag":122,"props":357,"children":358},{"style":129},[359],{"type":49,"value":152},{"type":43,"tag":122,"props":361,"children":363},{"style":362},"--shiki-light:#E2931D;--shiki-default:#FFCB6B;--shiki-dark:#FFCB6B",[364],{"type":49,"value":365},"seen_products",{"type":43,"tag":122,"props":367,"children":368},{"style":129},[369],{"type":49,"value":152},{"type":43,"tag":122,"props":371,"children":372},{"style":129},[373],{"type":49,"value":157},{"type":43,"tag":122,"props":375,"children":376},{"style":129},[377],{"type":49,"value":273},{"type":43,"tag":122,"props":379,"children":380},{"style":129},[381],{"type":49,"value":152},{"type":43,"tag":122,"props":383,"children":384},{"style":165},[385],{"type":49,"value":282},{"type":43,"tag":122,"props":387,"children":388},{"style":129},[389],{"type":49,"value":152},{"type":43,"tag":122,"props":391,"children":392},{"style":129},[393],{"type":49,"value":394},"],",{"type":43,"tag":122,"props":396,"children":397},{"style":129},[398],{"type":49,"value":162},{"type":43,"tag":122,"props":400,"children":401},{"style":362},[402],{"type":49,"value":403},"cart",{"type":43,"tag":122,"props":405,"children":406},{"style":129},[407],{"type":49,"value":152},{"type":43,"tag":122,"props":409,"children":410},{"style":129},[411],{"type":49,"value":157},{"type":43,"tag":122,"props":413,"children":414},{"style":129},[415],{"type":49,"value":273},{"type":43,"tag":122,"props":417,"children":419},{"style":418},"--shiki-light:#90A4AE;--shiki-default:#EEFFFF;--shiki-dark:#BABED8",[420],{"type":49,"value":282},{"type":43,"tag":122,"props":422,"children":423},{"style":129},[424],{"type":49,"value":394},{"type":43,"tag":122,"props":426,"children":427},{"style":129},[428],{"type":49,"value":162},{"type":43,"tag":122,"props":430,"children":431},{"style":362},[432],{"type":49,"value":433},"memory",{"type":43,"tag":122,"props":435,"children":436},{"style":129},[437],{"type":49,"value":152},{"type":43,"tag":122,"props":439,"children":440},{"style":129},[441],{"type":49,"value":157},{"type":43,"tag":122,"props":443,"children":444},{"style":129},[445],{"type":49,"value":273},{"type":43,"tag":122,"props":447,"children":448},{"style":418},[449],{"type":49,"value":282},{"type":43,"tag":122,"props":451,"children":452},{"style":129},[453],{"type":49,"value":394},{"type":43,"tag":122,"props":455,"children":456},{"style":129},[457],{"type":49,"value":162},{"type":43,"tag":122,"props":459,"children":460},{"style":362},[461],{"type":49,"value":462},"staged_changes",{"type":43,"tag":122,"props":464,"children":465},{"style":129},[466],{"type":49,"value":152},{"type":43,"tag":122,"props":468,"children":469},{"style":129},[470],{"type":49,"value":157},{"type":43,"tag":122,"props":472,"children":473},{"style":129},[474],{"type":49,"value":273},{"type":43,"tag":122,"props":476,"children":477},{"style":418},[478],{"type":49,"value":282},{"type":43,"tag":122,"props":480,"children":481},{"style":129},[482],{"type":49,"value":483},"]},\n",{"type":43,"tag":122,"props":485,"children":487},{"class":124,"line":486},6,[488,492,497,501,505,509,513,518,522,526,530,534,538],{"type":43,"tag":122,"props":489,"children":490},{"style":129},[491],{"type":49,"value":141},{"type":43,"tag":122,"props":493,"children":494},{"style":144},[495],{"type":49,"value":496},"turns",{"type":43,"tag":122,"props":498,"children":499},{"style":129},[500],{"type":49,"value":152},{"type":43,"tag":122,"props":502,"children":503},{"style":129},[504],{"type":49,"value":157},{"type":43,"tag":122,"props":506,"children":507},{"style":129},[508],{"type":49,"value":273},{"type":43,"tag":122,"props":510,"children":511},{"style":129},[512],{"type":49,"value":152},{"type":43,"tag":122,"props":514,"children":515},{"style":165},[516],{"type":49,"value":517},"\u003Cthe customer's or operator's message>",{"type":43,"tag":122,"props":519,"children":520},{"style":129},[521],{"type":49,"value":152},{"type":43,"tag":122,"props":523,"children":524},{"style":129},[525],{"type":49,"value":216},{"type":43,"tag":122,"props":527,"children":528},{"style":129},[529],{"type":49,"value":162},{"type":43,"tag":122,"props":531,"children":532},{"style":165},[533],{"type":49,"value":282},{"type":43,"tag":122,"props":535,"children":536},{"style":129},[537],{"type":49,"value":152},{"type":43,"tag":122,"props":539,"children":540},{"style":129},[541],{"type":49,"value":291},{"type":43,"tag":122,"props":543,"children":545},{"class":124,"line":544},7,[546,550,555,559,563],{"type":43,"tag":122,"props":547,"children":548},{"style":129},[549],{"type":49,"value":141},{"type":43,"tag":122,"props":551,"children":552},{"style":144},[553],{"type":49,"value":554},"expected",{"type":43,"tag":122,"props":556,"children":557},{"style":129},[558],{"type":49,"value":152},{"type":43,"tag":122,"props":560,"children":561},{"style":129},[562],{"type":49,"value":157},{"type":43,"tag":122,"props":564,"children":565},{"style":129},[566],{"type":49,"value":567}," {\n",{"type":43,"tag":122,"props":569,"children":571},{"class":124,"line":570},8,[572,577,582,586,590,594,598,602,606,610,615,620,624,628,632,636,640,644,648,653,658,662,666,670,674,678,682],{"type":43,"tag":122,"props":573,"children":574},{"style":129},[575],{"type":49,"value":576},"    \"",{"type":43,"tag":122,"props":578,"children":579},{"style":362},[580],{"type":49,"value":581},"calls_tool",{"type":43,"tag":122,"props":583,"children":584},{"style":129},[585],{"type":49,"value":152},{"type":43,"tag":122,"props":587,"children":588},{"style":129},[589],{"type":49,"value":157},{"type":43,"tag":122,"props":591,"children":592},{"style":129},[593],{"type":49,"value":273},{"type":43,"tag":122,"props":595,"children":596},{"style":129},[597],{"type":49,"value":152},{"type":43,"tag":122,"props":599,"children":600},{"style":165},[601],{"type":49,"value":282},{"type":43,"tag":122,"props":603,"children":604},{"style":129},[605],{"type":49,"value":152},{"type":43,"tag":122,"props":607,"children":608},{"style":129},[609],{"type":49,"value":394},{"type":43,"tag":122,"props":611,"children":612},{"style":129},[613],{"type":49,"value":614},"            \"",{"type":43,"tag":122,"props":616,"children":617},{"style":362},[618],{"type":49,"value":619},"calls_one_of",{"type":43,"tag":122,"props":621,"children":622},{"style":129},[623],{"type":49,"value":152},{"type":43,"tag":122,"props":625,"children":626},{"style":129},[627],{"type":49,"value":157},{"type":43,"tag":122,"props":629,"children":630},{"style":129},[631],{"type":49,"value":273},{"type":43,"tag":122,"props":633,"children":634},{"style":129},[635],{"type":49,"value":152},{"type":43,"tag":122,"props":637,"children":638},{"style":165},[639],{"type":49,"value":282},{"type":43,"tag":122,"props":641,"children":642},{"style":129},[643],{"type":49,"value":152},{"type":43,"tag":122,"props":645,"children":646},{"style":129},[647],{"type":49,"value":394},{"type":43,"tag":122,"props":649,"children":650},{"style":129},[651],{"type":49,"value":652},"          \"",{"type":43,"tag":122,"props":654,"children":655},{"style":362},[656],{"type":49,"value":657},"never_calls",{"type":43,"tag":122,"props":659,"children":660},{"style":129},[661],{"type":49,"value":152},{"type":43,"tag":122,"props":663,"children":664},{"style":129},[665],{"type":49,"value":157},{"type":43,"tag":122,"props":667,"children":668},{"style":129},[669],{"type":49,"value":273},{"type":43,"tag":122,"props":671,"children":672},{"style":129},[673],{"type":49,"value":152},{"type":43,"tag":122,"props":675,"children":676},{"style":165},[677],{"type":49,"value":282},{"type":43,"tag":122,"props":679,"children":680},{"style":129},[681],{"type":49,"value":152},{"type":43,"tag":122,"props":683,"children":684},{"style":129},[685],{"type":49,"value":291},{"type":43,"tag":122,"props":687,"children":689},{"class":124,"line":688},9,[690,694,699,703,707,711,715,719,723,728,733,737,741,745,749,753],{"type":43,"tag":122,"props":691,"children":692},{"style":129},[693],{"type":49,"value":576},{"type":43,"tag":122,"props":695,"children":696},{"style":362},[697],{"type":49,"value":698},"first_tool",{"type":43,"tag":122,"props":700,"children":701},{"style":129},[702],{"type":49,"value":152},{"type":43,"tag":122,"props":704,"children":705},{"style":129},[706],{"type":49,"value":157},{"type":43,"tag":122,"props":708,"children":709},{"style":129},[710],{"type":49,"value":162},{"type":43,"tag":122,"props":712,"children":713},{"style":165},[714],{"type":49,"value":282},{"type":43,"tag":122,"props":716,"children":717},{"style":129},[718],{"type":49,"value":152},{"type":43,"tag":122,"props":720,"children":721},{"style":129},[722],{"type":49,"value":216},{"type":43,"tag":122,"props":724,"children":725},{"style":129},[726],{"type":49,"value":727},"              \"",{"type":43,"tag":122,"props":729,"children":730},{"style":362},[731],{"type":49,"value":732},"first_tool_not",{"type":43,"tag":122,"props":734,"children":735},{"style":129},[736],{"type":49,"value":152},{"type":43,"tag":122,"props":738,"children":739},{"style":129},[740],{"type":49,"value":157},{"type":43,"tag":122,"props":742,"children":743},{"style":129},[744],{"type":49,"value":162},{"type":43,"tag":122,"props":746,"children":747},{"style":165},[748],{"type":49,"value":282},{"type":43,"tag":122,"props":750,"children":751},{"style":129},[752],{"type":49,"value":152},{"type":43,"tag":122,"props":754,"children":755},{"style":129},[756],{"type":49,"value":177},{"type":43,"tag":122,"props":758,"children":760},{"class":124,"line":759},10,[761,765,770,774,778,782,786,790,794,798,803,808,812,816],{"type":43,"tag":122,"props":762,"children":763},{"style":129},[764],{"type":49,"value":576},{"type":43,"tag":122,"props":766,"children":767},{"style":362},[768],{"type":49,"value":769},"ui_components",{"type":43,"tag":122,"props":771,"children":772},{"style":129},[773],{"type":49,"value":152},{"type":43,"tag":122,"props":775,"children":776},{"style":129},[777],{"type":49,"value":157},{"type":43,"tag":122,"props":779,"children":780},{"style":129},[781],{"type":49,"value":273},{"type":43,"tag":122,"props":783,"children":784},{"style":129},[785],{"type":49,"value":152},{"type":43,"tag":122,"props":787,"children":788},{"style":165},[789],{"type":49,"value":282},{"type":43,"tag":122,"props":791,"children":792},{"style":129},[793],{"type":49,"value":152},{"type":43,"tag":122,"props":795,"children":796},{"style":129},[797],{"type":49,"value":394},{"type":43,"tag":122,"props":799,"children":800},{"style":129},[801],{"type":49,"value":802},"         \"",{"type":43,"tag":122,"props":804,"children":805},{"style":362},[806],{"type":49,"value":807},"no_ui",{"type":43,"tag":122,"props":809,"children":810},{"style":129},[811],{"type":49,"value":152},{"type":43,"tag":122,"props":813,"children":814},{"style":129},[815],{"type":49,"value":157},{"type":43,"tag":122,"props":817,"children":818},{"style":129},[819],{"type":49,"value":820}," true,\n",{"type":43,"tag":122,"props":822,"children":824},{"class":124,"line":823},11,[825,829,834,838,842,846,850,854,858,862,866,871,875,879,885,889,894,899,903,907,911,915,919,923],{"type":43,"tag":122,"props":826,"children":827},{"style":129},[828],{"type":49,"value":576},{"type":43,"tag":122,"props":830,"children":831},{"style":362},[832],{"type":49,"value":833},"cart_contains",{"type":43,"tag":122,"props":835,"children":836},{"style":129},[837],{"type":49,"value":152},{"type":43,"tag":122,"props":839,"children":840},{"style":129},[841],{"type":49,"value":157},{"type":43,"tag":122,"props":843,"children":844},{"style":129},[845],{"type":49,"value":273},{"type":43,"tag":122,"props":847,"children":848},{"style":129},[849],{"type":49,"value":152},{"type":43,"tag":122,"props":851,"children":852},{"style":165},[853],{"type":49,"value":282},{"type":43,"tag":122,"props":855,"children":856},{"style":129},[857],{"type":49,"value":152},{"type":43,"tag":122,"props":859,"children":860},{"style":129},[861],{"type":49,"value":394},{"type":43,"tag":122,"props":863,"children":864},{"style":129},[865],{"type":49,"value":802},{"type":43,"tag":122,"props":867,"children":868},{"style":362},[869],{"type":49,"value":870},"cart_item_count",{"type":43,"tag":122,"props":872,"children":873},{"style":129},[874],{"type":49,"value":152},{"type":43,"tag":122,"props":876,"children":877},{"style":129},[878],{"type":49,"value":157},{"type":43,"tag":122,"props":880,"children":882},{"style":881},"--shiki-light:#F76D47;--shiki-default:#F78C6C;--shiki-dark:#F78C6C",[883],{"type":49,"value":884}," 0",{"type":43,"tag":122,"props":886,"children":887},{"style":129},[888],{"type":49,"value":216},{"type":43,"tag":122,"props":890,"children":891},{"style":129},[892],{"type":49,"value":893},"             \"",{"type":43,"tag":122,"props":895,"children":896},{"style":362},[897],{"type":49,"value":898},"cart_not_contains",{"type":43,"tag":122,"props":900,"children":901},{"style":129},[902],{"type":49,"value":152},{"type":43,"tag":122,"props":904,"children":905},{"style":129},[906],{"type":49,"value":157},{"type":43,"tag":122,"props":908,"children":909},{"style":129},[910],{"type":49,"value":273},{"type":43,"tag":122,"props":912,"children":913},{"style":129},[914],{"type":49,"value":152},{"type":43,"tag":122,"props":916,"children":917},{"style":165},[918],{"type":49,"value":282},{"type":43,"tag":122,"props":920,"children":921},{"style":129},[922],{"type":49,"value":152},{"type":43,"tag":122,"props":924,"children":925},{"style":129},[926],{"type":49,"value":291},{"type":43,"tag":122,"props":928,"children":930},{"class":124,"line":929},12,[931,935,940,944,948,952,956,960,964,968,972,977,981,985],{"type":43,"tag":122,"props":932,"children":933},{"style":129},[934],{"type":49,"value":576},{"type":43,"tag":122,"props":936,"children":937},{"style":362},[938],{"type":49,"value":939},"staged_change_kinds",{"type":43,"tag":122,"props":941,"children":942},{"style":129},[943],{"type":49,"value":152},{"type":43,"tag":122,"props":945,"children":946},{"style":129},[947],{"type":49,"value":157},{"type":43,"tag":122,"props":949,"children":950},{"style":129},[951],{"type":49,"value":273},{"type":43,"tag":122,"props":953,"children":954},{"style":129},[955],{"type":49,"value":152},{"type":43,"tag":122,"props":957,"children":958},{"style":165},[959],{"type":49,"value":282},{"type":43,"tag":122,"props":961,"children":962},{"style":129},[963],{"type":49,"value":152},{"type":43,"tag":122,"props":965,"children":966},{"style":129},[967],{"type":49,"value":394},{"type":43,"tag":122,"props":969,"children":970},{"style":129},[971],{"type":49,"value":221},{"type":43,"tag":122,"props":973,"children":974},{"style":362},[975],{"type":49,"value":976},"no_applied_changes",{"type":43,"tag":122,"props":978,"children":979},{"style":129},[980],{"type":49,"value":152},{"type":43,"tag":122,"props":982,"children":983},{"style":129},[984],{"type":49,"value":157},{"type":43,"tag":122,"props":986,"children":987},{"style":129},[988],{"type":49,"value":820},{"type":43,"tag":122,"props":990,"children":992},{"class":124,"line":991},13,[993,997,1002,1006,1010,1014,1018,1022,1026,1030,1035,1040,1044,1048,1052,1056,1060,1064],{"type":43,"tag":122,"props":994,"children":995},{"style":129},[996],{"type":49,"value":576},{"type":43,"tag":122,"props":998,"children":999},{"style":362},[1000],{"type":49,"value":1001},"memory_contains",{"type":43,"tag":122,"props":1003,"children":1004},{"style":129},[1005],{"type":49,"value":152},{"type":43,"tag":122,"props":1007,"children":1008},{"style":129},[1009],{"type":49,"value":157},{"type":43,"tag":122,"props":1011,"children":1012},{"style":129},[1013],{"type":49,"value":273},{"type":43,"tag":122,"props":1015,"children":1016},{"style":129},[1017],{"type":49,"value":152},{"type":43,"tag":122,"props":1019,"children":1020},{"style":165},[1021],{"type":49,"value":282},{"type":43,"tag":122,"props":1023,"children":1024},{"style":129},[1025],{"type":49,"value":152},{"type":43,"tag":122,"props":1027,"children":1028},{"style":129},[1029],{"type":49,"value":394},{"type":43,"tag":122,"props":1031,"children":1032},{"style":129},[1033],{"type":49,"value":1034},"       \"",{"type":43,"tag":122,"props":1036,"children":1037},{"style":362},[1038],{"type":49,"value":1039},"memory_not_contains",{"type":43,"tag":122,"props":1041,"children":1042},{"style":129},[1043],{"type":49,"value":152},{"type":43,"tag":122,"props":1045,"children":1046},{"style":129},[1047],{"type":49,"value":157},{"type":43,"tag":122,"props":1049,"children":1050},{"style":129},[1051],{"type":49,"value":273},{"type":43,"tag":122,"props":1053,"children":1054},{"style":129},[1055],{"type":49,"value":152},{"type":43,"tag":122,"props":1057,"children":1058},{"style":165},[1059],{"type":49,"value":282},{"type":43,"tag":122,"props":1061,"children":1062},{"style":129},[1063],{"type":49,"value":152},{"type":43,"tag":122,"props":1065,"children":1066},{"style":129},[1067],{"type":49,"value":291},{"type":43,"tag":122,"props":1069,"children":1071},{"class":124,"line":1070},14,[1072,1076,1081,1085,1089,1093,1097,1101,1105,1109,1114,1118,1122,1126,1130,1134,1138,1143,1148,1152,1156],{"type":43,"tag":122,"props":1073,"children":1074},{"style":129},[1075],{"type":49,"value":576},{"type":43,"tag":122,"props":1077,"children":1078},{"style":362},[1079],{"type":49,"value":1080},"skill_loaded",{"type":43,"tag":122,"props":1082,"children":1083},{"style":129},[1084],{"type":49,"value":152},{"type":43,"tag":122,"props":1086,"children":1087},{"style":129},[1088],{"type":49,"value":157},{"type":43,"tag":122,"props":1090,"children":1091},{"style":129},[1092],{"type":49,"value":162},{"type":43,"tag":122,"props":1094,"children":1095},{"style":165},[1096],{"type":49,"value":282},{"type":43,"tag":122,"props":1098,"children":1099},{"style":129},[1100],{"type":49,"value":152},{"type":43,"tag":122,"props":1102,"children":1103},{"style":129},[1104],{"type":49,"value":216},{"type":43,"tag":122,"props":1106,"children":1107},{"style":129},[1108],{"type":49,"value":614},{"type":43,"tag":122,"props":1110,"children":1111},{"style":362},[1112],{"type":49,"value":1113},"skill_not_loaded",{"type":43,"tag":122,"props":1115,"children":1116},{"style":129},[1117],{"type":49,"value":152},{"type":43,"tag":122,"props":1119,"children":1120},{"style":129},[1121],{"type":49,"value":157},{"type":43,"tag":122,"props":1123,"children":1124},{"style":129},[1125],{"type":49,"value":162},{"type":43,"tag":122,"props":1127,"children":1128},{"style":165},[1129],{"type":49,"value":282},{"type":43,"tag":122,"props":1131,"children":1132},{"style":129},[1133],{"type":49,"value":152},{"type":43,"tag":122,"props":1135,"children":1136},{"style":129},[1137],{"type":49,"value":216},{"type":43,"tag":122,"props":1139,"children":1140},{"style":129},[1141],{"type":49,"value":1142},"        \"",{"type":43,"tag":122,"props":1144,"children":1145},{"style":362},[1146],{"type":49,"value":1147},"no_skill_load",{"type":43,"tag":122,"props":1149,"children":1150},{"style":129},[1151],{"type":49,"value":152},{"type":43,"tag":122,"props":1153,"children":1154},{"style":129},[1155],{"type":49,"value":157},{"type":43,"tag":122,"props":1157,"children":1158},{"style":129},[1159],{"type":49,"value":820},{"type":43,"tag":122,"props":1161,"children":1163},{"class":124,"line":1162},15,[1164,1168,1173,1177,1181,1185,1189,1193,1197,1201,1205,1210,1214,1218,1222,1226,1230,1234,1238,1243,1248,1252,1256,1260],{"type":43,"tag":122,"props":1165,"children":1166},{"style":129},[1167],{"type":49,"value":576},{"type":43,"tag":122,"props":1169,"children":1170},{"style":362},[1171],{"type":49,"value":1172},"reply_includes",{"type":43,"tag":122,"props":1174,"children":1175},{"style":129},[1176],{"type":49,"value":152},{"type":43,"tag":122,"props":1178,"children":1179},{"style":129},[1180],{"type":49,"value":157},{"type":43,"tag":122,"props":1182,"children":1183},{"style":129},[1184],{"type":49,"value":273},{"type":43,"tag":122,"props":1186,"children":1187},{"style":129},[1188],{"type":49,"value":152},{"type":43,"tag":122,"props":1190,"children":1191},{"style":165},[1192],{"type":49,"value":282},{"type":43,"tag":122,"props":1194,"children":1195},{"style":129},[1196],{"type":49,"value":152},{"type":43,"tag":122,"props":1198,"children":1199},{"style":129},[1200],{"type":49,"value":394},{"type":43,"tag":122,"props":1202,"children":1203},{"style":129},[1204],{"type":49,"value":1142},{"type":43,"tag":122,"props":1206,"children":1207},{"style":362},[1208],{"type":49,"value":1209},"reply_omits",{"type":43,"tag":122,"props":1211,"children":1212},{"style":129},[1213],{"type":49,"value":152},{"type":43,"tag":122,"props":1215,"children":1216},{"style":129},[1217],{"type":49,"value":157},{"type":43,"tag":122,"props":1219,"children":1220},{"style":129},[1221],{"type":49,"value":273},{"type":43,"tag":122,"props":1223,"children":1224},{"style":129},[1225],{"type":49,"value":152},{"type":43,"tag":122,"props":1227,"children":1228},{"style":165},[1229],{"type":49,"value":282},{"type":43,"tag":122,"props":1231,"children":1232},{"style":129},[1233],{"type":49,"value":152},{"type":43,"tag":122,"props":1235,"children":1236},{"style":129},[1237],{"type":49,"value":394},{"type":43,"tag":122,"props":1239,"children":1240},{"style":129},[1241],{"type":49,"value":1242},"           \"",{"type":43,"tag":122,"props":1244,"children":1245},{"style":362},[1246],{"type":49,"value":1247},"max_tool_calls",{"type":43,"tag":122,"props":1249,"children":1250},{"style":129},[1251],{"type":49,"value":152},{"type":43,"tag":122,"props":1253,"children":1254},{"style":129},[1255],{"type":49,"value":157},{"type":43,"tag":122,"props":1257,"children":1258},{"style":881},[1259],{"type":49,"value":884},{"type":43,"tag":122,"props":1261,"children":1262},{"style":129},[1263],{"type":49,"value":177},{"type":43,"tag":122,"props":1265,"children":1267},{"class":124,"line":1266},16,[1268,1272,1277,1281,1285,1289,1294],{"type":43,"tag":122,"props":1269,"children":1270},{"style":129},[1271],{"type":49,"value":576},{"type":43,"tag":122,"props":1273,"children":1274},{"style":362},[1275],{"type":49,"value":1276},"rubric",{"type":43,"tag":122,"props":1278,"children":1279},{"style":129},[1280],{"type":49,"value":152},{"type":43,"tag":122,"props":1282,"children":1283},{"style":129},[1284],{"type":49,"value":157},{"type":43,"tag":122,"props":1286,"children":1287},{"style":129},[1288],{"type":49,"value":162},{"type":43,"tag":122,"props":1290,"children":1291},{"style":165},[1292],{"type":49,"value":1293},"PASS if \u003Ccondition>. FAIL if \u003Ccondition>.",{"type":43,"tag":122,"props":1295,"children":1296},{"style":129},[1297],{"type":49,"value":1298},"\"\n",{"type":43,"tag":122,"props":1300,"children":1302},{"class":124,"line":1301},17,[1303],{"type":43,"tag":122,"props":1304,"children":1305},{"style":129},[1306],{"type":49,"value":1307},"  },\n",{"type":43,"tag":122,"props":1309,"children":1311},{"class":124,"line":1310},18,[1312,1316,1321,1325,1329,1333,1338],{"type":43,"tag":122,"props":1313,"children":1314},{"style":129},[1315],{"type":49,"value":141},{"type":43,"tag":122,"props":1317,"children":1318},{"style":144},[1319],{"type":49,"value":1320},"notes",{"type":43,"tag":122,"props":1322,"children":1323},{"style":129},[1324],{"type":49,"value":152},{"type":43,"tag":122,"props":1326,"children":1327},{"style":129},[1328],{"type":49,"value":157},{"type":43,"tag":122,"props":1330,"children":1331},{"style":129},[1332],{"type":49,"value":162},{"type":43,"tag":122,"props":1334,"children":1335},{"style":165},[1336],{"type":49,"value":1337},"\u003Cwhat the case pins and the fixture fact that decides it>",{"type":43,"tag":122,"props":1339,"children":1340},{"style":129},[1341],{"type":49,"value":1298},{"type":43,"tag":122,"props":1343,"children":1345},{"class":124,"line":1344},19,[1346],{"type":43,"tag":122,"props":1347,"children":1348},{"style":129},[1349],{"type":49,"value":1350},"}\n",{"type":43,"tag":52,"props":1352,"children":1353},{},[1354,1359,1361,1366,1368,1373,1375,1380,1382,1387,1389,1394,1395,1400,1402,1407],{"type":43,"tag":58,"props":1355,"children":1357},{"className":1356},[],[1358],{"type":49,"value":342},{"type":49,"value":1360}," is the precondition; ",{"type":43,"tag":58,"props":1362,"children":1364},{"className":1363},[],[1365],{"type":49,"value":496},{"type":49,"value":1367}," is one message unless the behavior under test is carrying state across turns;\n",{"type":43,"tag":58,"props":1369,"children":1371},{"className":1370},[],[1372],{"type":49,"value":554},{"type":49,"value":1374}," holds only the keys the case is about. Ids in ",{"type":43,"tag":58,"props":1376,"children":1378},{"className":1377},[],[1379],{"type":49,"value":342},{"type":49,"value":1381}," and ",{"type":43,"tag":58,"props":1383,"children":1385},{"className":1384},[],[1386],{"type":49,"value":554},{"type":49,"value":1388}," are real ids from your fixtures.\n",{"type":43,"tag":58,"props":1390,"children":1392},{"className":1391},[],[1393],{"type":49,"value":190},{"type":49,"value":1381},{"type":43,"tag":58,"props":1396,"children":1398},{"className":1397},[],[1399],{"type":49,"value":226},{"type":49,"value":1401}," let a report say which failures matter; a case that cannot run yet carries ",{"type":43,"tag":58,"props":1403,"children":1405},{"className":1404},[],[1406],{"type":49,"value":304},{"type":49,"value":1408},"\nwith its reason rather than being deleted.",{"type":43,"tag":91,"props":1410,"children":1412},{"id":1411},"authoring-rules",[1413],{"type":49,"value":1414},"Authoring rules",{"type":43,"tag":1416,"props":1417,"children":1418},"ul",{},[1419,1425,1430,1435,1461,1466,1493],{"type":43,"tag":1420,"props":1421,"children":1422},"li",{},[1423],{"type":49,"value":1424},"Preconditions go in injected state (the products already seen, the cart, the memory facts, the staged queue), which\nthe runner loads into the session state and the memory store before the turn. Earlier turns are for state the\nbehavior itself carries, and for nothing else.",{"type":43,"tag":1420,"props":1426,"children":1427},{},[1428],{"type":49,"value":1429},"Every positive has a negative: for each case that asserts a component, a skill load, a gate, a disclosure, or a\nmemory write, a case in the same niche asserts its absence. A refusal case has a should-serve counterpart.",{"type":43,"tag":1420,"props":1431,"children":1432},{},[1433],{"type":49,"value":1434},"Memory is three cases: a remark worth keeping is written; an identifier or excluded content is refused (nothing\nstored, no error to the person); a stored fact changes the next session's pick.",{"type":43,"tag":1420,"props":1436,"children":1437},{},[1438,1440,1445,1447,1452,1454,1459],{"type":49,"value":1439},"Grade the final tool arguments and the state they produced: the ids on the last presentation call, the fields on\nthe staged change, the cart and the memory store after the turn. The reply's wording is graded only for strings\nthat must or must not appear. Which route the agent took (",{"type":43,"tag":58,"props":1441,"children":1443},{"className":1442},[],[1444],{"type":49,"value":1080},{"type":49,"value":1446},", one named component, ",{"type":43,"tag":58,"props":1448,"children":1450},{"className":1449},[],[1451],{"type":49,"value":657},{"type":49,"value":1453}," on a\npresentation tool) is asserted only where the route is the behavior (a grounding read first, a write that must\nnever happen); elsewhere ",{"type":43,"tag":58,"props":1455,"children":1457},{"className":1456},[],[1458],{"type":49,"value":619},{"type":49,"value":1460}," names the acceptable set. When a live run takes a route the case did not\nexpect and the answer was right, widen the case to the acceptable set; do not re-pin it to the route observed.",{"type":43,"tag":1420,"props":1462,"children":1463},{},[1464],{"type":49,"value":1465},"A rubric is one PASS condition and one FAIL condition that no response satisfies both of; it names the fixture\nfact that decides it (the updated delivery date, the price today); variants you accept are written into it; it\nsays nothing about tone, length, or the order components appear in.",{"type":43,"tag":1420,"props":1467,"children":1468},{},[1469,1471,1476,1478,1483,1485,1491],{"type":49,"value":1470},"A turn that mentions health, a one-off errand, or hostile content asserts the memory end-state\n(",{"type":43,"tag":58,"props":1472,"children":1474},{"className":1473},[],[1475],{"type":49,"value":1039},{"type":49,"value":1477},", or ",{"type":43,"tag":58,"props":1479,"children":1481},{"className":1480},[],[1482],{"type":49,"value":657},{"type":49,"value":1484}," on ",{"type":43,"tag":58,"props":1486,"children":1488},{"className":1487},[],[1489],{"type":49,"value":1490},"save_memory",{"type":49,"value":1492},"). A case where a stored fact should change the pick uses a\nquery whose results contain both the item the fact favors and the one it rules out; run the search before writing\nthe case.",{"type":43,"tag":1420,"props":1494,"children":1495},{},[1496,1501,1503,1509],{"type":43,"tag":58,"props":1497,"children":1499},{"className":1498},[],[1500],{"type":49,"value":1247},{"type":49,"value":1502}," is set from what a well-behaved agent needs; a multi-item request fans out several searches in one round, and the ",{"type":43,"tag":58,"props":1504,"children":1506},{"className":1505},[],[1507],{"type":49,"value":1508},"present_suggestions",{"type":49,"value":1510}," call that ends the turn is not counted.",{"type":43,"tag":91,"props":1512,"children":1514},{"id":1513},"scorers",[1515],{"type":49,"value":1516},"Scorers",{"type":43,"tag":1416,"props":1518,"children":1519},{},[1520,1604,1614],{"type":43,"tag":1420,"props":1521,"children":1522},{},[1523,1525,1531,1533,1539,1541,1547,1549,1555,1557,1563,1565,1571,1573,1579,1581,1587,1589,1595,1597,1602],{"type":49,"value":1524},"Code graders read the events a turn yields (",{"type":43,"tag":58,"props":1526,"children":1528},{"className":1527},[],[1529],{"type":49,"value":1530},"commerce_common\u002Fstreaming.py",{"type":49,"value":1532},"): ",{"type":43,"tag":58,"props":1534,"children":1536},{"className":1535},[],[1537],{"type":49,"value":1538},"tool_call",{"type":49,"value":1540}," names and arguments,\n",{"type":43,"tag":58,"props":1542,"children":1544},{"className":1543},[],[1545],{"type":49,"value":1546},"tool_result",{"type":49,"value":1548}," with ",{"type":43,"tag":58,"props":1550,"children":1552},{"className":1551},[],[1553],{"type":49,"value":1554},"status",{"type":49,"value":1556}," ",{"type":43,"tag":58,"props":1558,"children":1560},{"className":1559},[],[1561],{"type":49,"value":1562},"blocked",{"type":49,"value":1564}," and the gate in ",{"type":43,"tag":58,"props":1566,"children":1568},{"className":1567},[],[1569],{"type":49,"value":1570},"reason",{"type":49,"value":1572},", ",{"type":43,"tag":58,"props":1574,"children":1576},{"className":1575},[],[1577],{"type":49,"value":1578},"ui",{"type":49,"value":1580}," component names, the last ",{"type":43,"tag":58,"props":1582,"children":1584},{"className":1583},[],[1585],{"type":49,"value":1586},"cart_update",{"type":49,"value":1588}," or\n",{"type":43,"tag":58,"props":1590,"children":1592},{"className":1591},[],[1593],{"type":49,"value":1594},"change_update",{"type":49,"value":1596},", and the reply text. Every key above except ",{"type":43,"tag":58,"props":1598,"children":1600},{"className":1599},[],[1601],{"type":49,"value":1276},{"type":49,"value":1603}," is a code grader.",{"type":43,"tag":1420,"props":1605,"children":1606},{},[1607,1612],{"type":43,"tag":58,"props":1608,"children":1610},{"className":1609},[],[1611],{"type":49,"value":1276},{"type":49,"value":1613}," goes to a judge. One judge call per dimension (budget respected, no invented availability, trade-off stated),\nreturning structured output with a verdict and a reason; the transcript is passed to it as quoted material, tool\nresults and component payloads included; when the transcript exceeds the judge's window, truncate from the start so\nthe graded turn survives, and record the truncation on the outcome. Pin the judge model at temperature zero; a change to the judge model or a rubric\ninvalidates every stored verdict scored with it, so the recording carries a fingerprint of both.",{"type":43,"tag":1420,"props":1615,"children":1616},{},[1617],{"type":49,"value":1618},"A judge reply that does not parse into a verdict is a judge failure on the case, kept apart from an agent failure.",{"type":43,"tag":91,"props":1620,"children":1622},{"id":1621},"run-pattern",[1623],{"type":49,"value":1624},"Run pattern",{"type":43,"tag":1626,"props":1627,"children":1628},"table",{},[1629,1653],{"type":43,"tag":1630,"props":1631,"children":1632},"thead",{},[1633],{"type":43,"tag":1634,"props":1635,"children":1636},"tr",{},[1637,1643,1648],{"type":43,"tag":1638,"props":1639,"children":1640},"th",{},[1641],{"type":49,"value":1642},"When",{"type":43,"tag":1638,"props":1644,"children":1645},{},[1646],{"type":49,"value":1647},"What runs",{"type":43,"tag":1638,"props":1649,"children":1650},{},[1651],{"type":49,"value":1652},"What decides",{"type":43,"tag":1654,"props":1655,"children":1656},"tbody",{},[1657,1676,1694,1712],{"type":43,"tag":1634,"props":1658,"children":1659},{},[1660,1666,1671],{"type":43,"tag":1661,"props":1662,"children":1663},"td",{},[1664],{"type":49,"value":1665},"Every merge to the agent, a skill, a tool description, or a fixture",{"type":43,"tag":1661,"props":1667,"children":1668},{},[1669],{"type":49,"value":1670},"The regression set",{"type":43,"tag":1661,"props":1672,"children":1673},{},[1674],{"type":49,"value":1675},"Each case over several trials; a pass threshold per set",{"type":43,"tag":1634,"props":1677,"children":1678},{},[1679,1684,1689],{"type":43,"tag":1661,"props":1680,"children":1681},{},[1682],{"type":49,"value":1683},"While changing one flow",{"type":43,"tag":1661,"props":1685,"children":1686},{},[1687],{"type":49,"value":1688},"That flow's targeted set",{"type":43,"tag":1661,"props":1690,"children":1691},{},[1692],{"type":49,"value":1693},"The failure set, read beside the previous run's as the baseline",{"type":43,"tag":1634,"props":1695,"children":1696},{},[1697,1702,1707],{"type":43,"tag":1661,"props":1698,"children":1699},{},[1700],{"type":49,"value":1701},"Choosing or upgrading a model (commerce-architecture's model fields)",{"type":43,"tag":1661,"props":1703,"children":1704},{},[1705],{"type":49,"value":1706},"Everything",{"type":43,"tag":1661,"props":1708,"children":1709},{},[1710],{"type":49,"value":1711},"The two failure sets side by side",{"type":43,"tag":1634,"props":1713,"children":1714},{},[1715,1720,1725],{"type":43,"tag":1661,"props":1716,"children":1717},{},[1718],{"type":49,"value":1719},"In production",{"type":43,"tag":1661,"props":1721,"children":1722},{},[1723],{"type":49,"value":1724},"A judged sample of live traffic against the same rubrics",{"type":43,"tag":1661,"props":1726,"children":1727},{},[1728],{"type":49,"value":1729},"Trend per dimension",{"type":43,"tag":52,"props":1731,"children":1732},{},[1733],{"type":49,"value":1734},"Diff failure sets; a topline moving a point between live runs is noise. A case that fails after a change means the\nchange broke the behavior or the case encoded a stale one; fix whichever it is and say which in the commit.",{"type":43,"tag":91,"props":1736,"children":1738},{"id":1737},"poisoned-fixtures",[1739],{"type":49,"value":1740},"Poisoned fixtures",{"type":43,"tag":52,"props":1742,"children":1743},{},[1744,1746,1751,1752,1757,1758,1763,1764,1769,1770,1775],{"type":49,"value":1745},"Listings, reviews, and messages carrying instructions live in eval-only fixtures the runner merges into the backend\nfor the run, under a third-party brand or seller; none of their ids appears in demo data, seeds, or captures. Each\nsuch case asserts the negative in code (",{"type":43,"tag":58,"props":1747,"children":1749},{"className":1748},[],[1750],{"type":49,"value":657},{"type":49,"value":1572},{"type":43,"tag":58,"props":1753,"children":1755},{"className":1754},[],[1756],{"type":49,"value":898},{"type":49,"value":1572},{"type":43,"tag":58,"props":1759,"children":1761},{"className":1760},[],[1762],{"type":49,"value":976},{"type":49,"value":177},{"type":43,"tag":58,"props":1765,"children":1767},{"className":1766},[],[1768],{"type":49,"value":1039},{"type":49,"value":1572},{"type":43,"tag":58,"props":1771,"children":1773},{"className":1772},[],[1774],{"type":49,"value":1209},{"type":49,"value":1776},"), and every vector it asserts is one the driven turn actually puts in front of\nthe model (a review is only read on a details call). Cover at least an instruction to write to the cart or stage a\nchange, one to remember something, and a false claim (a code, a guarantee). The should-serve counterpart is a\nseparate benign eval-only listing in the same niche, so an agent that refuses everything fails it.",{"type":43,"tag":1778,"props":1779,"children":1780},"style",{},[1781],{"type":49,"value":1782},"html .light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html.light .shiki span {color: var(--shiki-light);background: var(--shiki-light-bg);font-style: var(--shiki-light-font-style);font-weight: var(--shiki-light-font-weight);text-decoration: var(--shiki-light-text-decoration);}html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}html.dark .shiki span {color: var(--shiki-dark);background: var(--shiki-dark-bg);font-style: var(--shiki-dark-font-style);font-weight: var(--shiki-dark-font-weight);text-decoration: var(--shiki-dark-text-decoration);}",{"items":1784,"total":1968},[1785,1801,1820,1834,1846,1863,1879,1890,1911,1931,1945,1960],{"slug":1786,"name":1786,"fn":1787,"description":1788,"org":1789,"tags":1790,"stars":1798,"repoUrl":1799,"updatedAt":1800},"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},[1791,1792,1795],{"name":9,"slug":8,"type":16},{"name":1793,"slug":1794,"type":16},"Documentation","documentation",{"name":1796,"slug":1797,"type":16},"Education","education",161831,"https:\u002F\u002Fgithub.com\u002Fanthropics\u002Fskills","2026-08-19T03:59:01.021254",{"slug":1802,"name":1802,"fn":1803,"description":1804,"org":1805,"tags":1806,"stars":1798,"repoUrl":1799,"updatedAt":1819},"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},[1807,1810,1813,1816],{"name":1808,"slug":1809,"type":16},"Creative","creative",{"name":1811,"slug":1812,"type":16},"Design","design",{"name":1814,"slug":1815,"type":16},"Generative Art","generative-art",{"name":1817,"slug":1818,"type":16},"JavaScript","javascript","2026-04-06T17:56:15.455818",{"slug":1821,"name":1821,"fn":1822,"description":1823,"org":1824,"tags":1825,"stars":1798,"repoUrl":1799,"updatedAt":1833},"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},[1826,1829,1830],{"name":1827,"slug":1828,"type":16},"Branding","branding",{"name":1811,"slug":1812,"type":16},{"name":1831,"slug":1832,"type":16},"Typography","typography","2026-04-06T17:56:05.042852",{"slug":1835,"name":1835,"fn":1836,"description":1837,"org":1838,"tags":1839,"stars":1798,"repoUrl":1799,"updatedAt":1845},"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},[1840,1841,1842],{"name":1808,"slug":1809,"type":16},{"name":1811,"slug":1812,"type":16},{"name":1843,"slug":1844,"type":16},"PDF","pdf","2026-04-06T17:56:03.794732",{"slug":1847,"name":1847,"fn":1848,"description":1849,"org":1850,"tags":1851,"stars":1798,"repoUrl":1799,"updatedAt":1862},"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},[1852,1853,1854,1857,1859],{"name":21,"slug":22,"type":16},{"name":9,"slug":8,"type":16},{"name":1855,"slug":1856,"type":16},"Anthropic SDK","anthropic-sdk",{"name":1858,"slug":1847,"type":16},"Claude API",{"name":1860,"slug":1861,"type":16},"LLM","llm","2026-09-04T07:28:24.898544",{"slug":1864,"name":1864,"fn":1865,"description":1866,"org":1867,"tags":1868,"stars":1798,"repoUrl":1799,"updatedAt":1878},"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},[1869,1872,1875],{"name":1870,"slug":1871,"type":16},"Coaching","coaching",{"name":1873,"slug":1874,"type":16},"Productivity","productivity",{"name":1876,"slug":1877,"type":16},"Strategy","strategy","2026-08-19T03:59:01.539281",{"slug":1880,"name":1880,"fn":1881,"description":1882,"org":1883,"tags":1884,"stars":1798,"repoUrl":1799,"updatedAt":1889},"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},[1885,1886],{"name":1793,"slug":1794,"type":16},{"name":1887,"slug":1888,"type":16},"Technical Writing","technical-writing","2026-04-06T17:56:14.18897",{"slug":1891,"name":1891,"fn":1892,"description":1893,"org":1894,"tags":1895,"stars":1798,"repoUrl":1799,"updatedAt":1910},"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},[1896,1899,1901,1904,1907],{"name":1897,"slug":1898,"type":16},"Documents","documents",{"name":1900,"slug":1891,"type":16},"DOCX",{"name":1902,"slug":1903,"type":16},"Office","office",{"name":1905,"slug":1906,"type":16},"Templates","templates",{"name":1908,"slug":1909,"type":16},"Word","word","2026-07-18T05:16:23.136271",{"slug":1912,"name":1912,"fn":1913,"description":1914,"org":1915,"tags":1916,"stars":1798,"repoUrl":1799,"updatedAt":1930},"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},[1917,1918,1921,1924,1927],{"name":1811,"slug":1812,"type":16},{"name":1919,"slug":1920,"type":16},"Frontend","frontend",{"name":1922,"slug":1923,"type":16},"React","react",{"name":1925,"slug":1926,"type":16},"Tailwind CSS","tailwind-css",{"name":1928,"slug":1929,"type":16},"UI Components","ui-components","2026-09-04T07:28:23.795756",{"slug":1932,"name":1932,"fn":1933,"description":1934,"org":1935,"tags":1936,"stars":1798,"repoUrl":1799,"updatedAt":1944},"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},[1937,1940,1941],{"name":1938,"slug":1939,"type":16},"Communications","communications",{"name":1905,"slug":1906,"type":16},{"name":1942,"slug":1943,"type":16},"Writing","writing","2026-04-06T17:56:20.695522",{"slug":1946,"name":1946,"fn":1947,"description":1948,"org":1949,"tags":1950,"stars":1798,"repoUrl":1799,"updatedAt":1959},"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},[1951,1952,1955,1956],{"name":21,"slug":22,"type":16},{"name":1953,"slug":1954,"type":16},"API Development","api-development",{"name":1860,"slug":1861,"type":16},{"name":1957,"slug":1958,"type":16},"MCP","mcp","2026-04-06T17:56:10.357665",{"slug":1844,"name":1844,"fn":1961,"description":1962,"org":1963,"tags":1964,"stars":1798,"repoUrl":1799,"updatedAt":1967},"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},[1965,1966],{"name":1897,"slug":1898,"type":16},{"name":1843,"slug":1844,"type":16},"2026-04-06T17:56:02.483316",521,{"items":1970,"total":1266},[1971,1988,2003,2010,2025,2040,2056],{"slug":1972,"name":1972,"fn":1973,"description":1974,"org":1975,"tags":1976,"stars":26,"repoUrl":27,"updatedAt":1987},"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},[1977,1980,1983,1984],{"name":1978,"slug":1979,"type":16},"Content Creation","content-creation",{"name":1981,"slug":1982,"type":16},"Content Strategy","content-strategy",{"name":18,"slug":19,"type":16},{"name":1985,"slug":1986,"type":16},"Marketing","marketing","2026-09-04T08:00:35.843558",{"slug":1989,"name":1989,"fn":1990,"description":1991,"org":1992,"tags":1993,"stars":26,"repoUrl":27,"updatedAt":2002},"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},[1994,1995,1998,1999],{"name":21,"slug":22,"type":16},{"name":1996,"slug":1997,"type":16},"Architecture","architecture",{"name":18,"slug":19,"type":16},{"name":2000,"slug":2001,"type":16},"Engineering","engineering","2026-09-04T08:00:39.152322",{"slug":4,"name":4,"fn":5,"description":6,"org":2004,"tags":2005,"stars":26,"repoUrl":27,"updatedAt":28},{"slug":8,"name":9,"logoUrl":10,"githubOrg":11},[2006,2007,2008,2009],{"name":21,"slug":22,"type":16},{"name":18,"slug":19,"type":16},{"name":14,"slug":15,"type":16},{"name":24,"slug":25,"type":16},{"slug":2011,"name":2011,"fn":2012,"description":2013,"org":2014,"tags":2015,"stars":26,"repoUrl":27,"updatedAt":2024},"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},[2016,2017,2018,2021],{"name":21,"slug":22,"type":16},{"name":18,"slug":19,"type":16},{"name":2019,"slug":2020,"type":16},"Management","management",{"name":2022,"slug":2023,"type":16},"Operations","operations","2026-09-04T08:00:35.464555",{"slug":2026,"name":2026,"fn":2027,"description":2028,"org":2029,"tags":2030,"stars":26,"repoUrl":27,"updatedAt":2039},"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},[2031,2032,2033,2036],{"name":21,"slug":22,"type":16},{"name":18,"slug":19,"type":16},{"name":2034,"slug":2035,"type":16},"Performance","performance",{"name":2037,"slug":2038,"type":16},"Prompt Engineering","prompt-engineering","2026-09-04T08:00:39.50364",{"slug":2041,"name":2041,"fn":2042,"description":2043,"org":2044,"tags":2045,"stars":26,"repoUrl":27,"updatedAt":2055},"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},[2046,2047,2050,2052],{"name":21,"slug":22,"type":16},{"name":2048,"slug":2049,"type":16},"Compliance","compliance",{"name":2051,"slug":433,"type":16},"Memory",{"name":2053,"slug":2054,"type":16},"Security","security","2026-09-04T08:00:42.609563",{"slug":2057,"name":2057,"fn":2058,"description":2059,"org":2060,"tags":2061,"stars":26,"repoUrl":27,"updatedAt":2066},"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},[2062,2063,2064,2065],{"name":1811,"slug":1812,"type":16},{"name":18,"slug":19,"type":16},{"name":1919,"slug":1920,"type":16},{"name":1928,"slug":1929,"type":16},"2026-09-04T08:00:38.801247"]