[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7BV5YV8MZ10RES3SD4SQV3E":3,"skill-same:01M7BV5YV8MZ10RES3SD4SQV3E":49},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":28,"commitSha":29,"skillDirectory":30,"forksCount":31,"repositoryHtmlUrl":32,"repositoryHomepageUrl":6,"ownerHtmlUrl":33,"ownerBlogUrl":6,"ownerDescription":34,"updatedAt":35,"install":36,"resources":48},"01M7BV5YV8MZ10RES3SD4SQV3E","control-demo-roast",null,"Adversarial pre-announcement review of a Forged with Foundry control demo -- run before linkedin-post-writer, not instead of it -- to catch overclaiming, unproven mechanisms, and stale estimates that would embarrass the team once a technical community actually tries the demo","doruit\u002Fforged-with-foundry","doruit","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F144165645?v=4","01M7BV5HKR7V6XSAZ13DBMXHW5","Douwe van de Ruit","01M7BV5HMDT1WMC9W91K0K08TR",2,"MIT","https:\u002F\u002Fgithub.com\u002Fdoruit\u002Fforged-with-foundry\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fcontrol-demo-roast\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: control-demo-roast\ndescription: \"Adversarial pre-announcement review of a Forged with Foundry control demo -- run before linkedin-post-writer, not instead of it -- to catch overclaiming, unproven mechanisms, and stale estimates that would embarrass the team once a technical community actually tries the demo\"\nlicense: MIT\nuser-invocable: true\nmetadata:\n  authors: \"Douwe van de Ruit\"\n  spec_version: \"1.0\"\n  last_updated: \"2026-09-25\"\n---\n\n# Control Demo Roast Skill\n\n## Overview\n\nA control can pass every automated test, satisfy `markdown-doc-quality`'s\nchecklist, and still not deserve a public announcement, because none of\nthat checks whether the thing being announced actually does what the\nannouncement will claim. This skill is a deliberately adversarial, skeptical\npass over one control's full state (README, ASSESSMENT, code, tests, and\nwhat was actually run live) run **before** `linkedin-post-writer` drafts\nanything, and before a maintainer marks a control `Implemented`\u002F`Validated`\nin a change intended to be announced. Its job is to find the exact\nobjection a sharp member of the tech community would raise within the first\nfive minutes of actually trying the demo, and surface it now instead of in\npublic.\n\nThis complements `markdown-doc-quality` rather than replacing it:\n`markdown-doc-quality` catches writing-quality problems (hedging,\ncontradictions, missing evidence) in the prose itself. This skill catches a\ndifferent class of problem: prose that is well-written, internally\nconsistent, and passes every check, but describes a system whose central\nclaim has not actually been observed to hold. Apply both; a control can\nfail this skill's checklist while passing every item in that one.\n\n## When to use\n\n- Before drafting a `linkedin-post-writer` announcement for any control.\n- Before changing a control's `README.md` Status line to `Implemented` or\n  `Validated` in a change that will be publicized.\n- When a user asks to \"roast,\" \"sanity-check,\" or \"review before announcing\"\n  a control demo.\n- Not needed for a routine internal change that isn't headed toward a\n  public announcement or a status change.\n\n## The checklist\n\nWork through every item against the control's actual current state --\nREADME, ASSESSMENT.md, the real code, the real tests, and what was actually\nrun live this session -- not against what the control was designed to do.\nCite the exact file, section, or line each finding comes from; a finding\nwith no location is not actionable.\n\n#### 1. Status vs. observed reality\n\nDoes `Implemented`\u002F`Validated` describe what has actually been *observed\nworking*, or only what has been coded, unit-tested, and designed to work?\nFind the control's own authoritative signal (the thing its Control contract\nnames as \"Evidence \u002F source\") and ask: has this specific signal, from the\nspecific real service call the control depends on, ever actually been\nobserved producing its real output in a live run -- not a mocked, stubbed,\nor synthetic substitute? If the demo has only ever produced the *fail-closed*\nor *cannot-evaluate* branch of its own decision logic in every real run to\ndate, that is a specific, nameable gap: the control's core mechanism is\nunproven, not merely \"not yet exhaustively documented.\" State this\nplainly and do not let confident prose elsewhere in the README soften it.\n\n#### 2. Self-fulfilling or trivial proof\n\nWhen the demo's evidence includes a deliberately misconfigured or\nlower-quality component \"catching\" a problem, ask what was actually proven:\nthat the *monitoring\u002Fdetection mechanism* caught a real, independently-scored\nissue, or only that a component built to misbehave, in fact misbehaves. The\nsecond is not evidence the control's actual detection machinery works --\nit is evidence the demo's synthetic fixture is internally consistent. Name\nthis distinction explicitly when the \"breach\" scenario's detection came from\nconstruction (the reviewer already knows which component will fail) rather\nthan from the control's own authoritative signal actually flagging it\nindependently.\n\n#### 3. Unsubstantiated causal or outcome claims\n\nFlag any claim that a mechanism \"improves,\" \"raises,\" \"reduces,\" or \"heals\"\nsome real-world outcome over time, even when hedged (\"can help,\" \"may\nraise\"), when no experiment, measurement, or before\u002Fafter comparison in the\nsame control backs the causal claim. A plausible-sounding UX idea (for\nexample, showing a self-reported confidence score to a user) is not the\nsame as a demonstrated mechanism, and confident framing (\"a self-healing\nmechanism\") should not outrun what was actually measured. Recommend either\nsoftening the claim to what is actually true (a hypothesis, a design\nintent) or citing the actual evidence if it exists.\n\n#### 4. Complexity and time-estimate drift\n\nCompare the README's stated `Learning level` and `Estimated time` against\nthe *current* actual complexity: count real IAM role assignments, external\nresources, undocumented prerequisites, and any external dependency with\nunknown or open-ended latency (an async service that has never yet returned\nits result, RBAC propagation delays hit during the same session's live\ntesting, etc.). An estimate written when the design was simpler and never\nrevisited after it grew is a stale claim, not a current one. Recommend a\nrevised, honest estimate or an explicit caveat about the open-ended\ndependency.\n\n#### 5. Tone risk when citing an upstream gap\n\nWhen the control's writeup includes \"found a bug in the vendor SDK\" or \"the\nvendor's documentation has a gap,\" check whether the overall narrative could\nread as blaming the platform for the demo's own core mechanism not working,\nespecially when that mechanism is still unproven (see item 1). A real,\nindependently-reproduced upstream finding is worth citing professionally and\nfactually; it reads very differently next to an honest \"and our own\ncontrol doesn't fully work yet either\" than next to confident\n`Implemented` framing that omits it. Recommend adjusting emphasis, not\nremoving a legitimate finding.\n\n#### 6. Test coverage vs. what a first-time user will actually witness\n\nCheck whether any automated test pins down the *specific outcome* a person\nrunning the documented demo commands today will actually see -- not just\nclean unit tests of internal functions with synthetic inputs. If the real,\ncurrent, only-ever-observed outcome of running the demo is a specific\ndecision (for example, always `cannot_evaluate`), and no test asserts that\noutcome against the real integration point (not a mocked one), flag this as\na coverage gap distinct from \"tests exist and pass\": the tests do not\nprotect the one behavior most likely to surprise a new user.\n\n#### 7. Borrowed assumptions never confirmed against the real integration\n\nFlag any data shape, schema, or contract \"confirmed\" against a *different*\nsystem, mode, or code path (a legacy mechanism, a different API, an earlier\ndesign) and then assumed, without confirmation, to hold for the *new*\nmechanism the control now claims to use. If the new integration's real\noutput has never actually been observed, say so explicitly rather than\nletting a schema borrowed from elsewhere read as verified.\n\n#### 8. Stale or unverified external citations\n\nFor every external link or issue reference presented as current evidence\n(a GitHub issue, a vendor doc), confirm it was actually checked recently\nenough to still be accurate -- not simply carried forward from when it was\nfirst found. A citation whose status (open\u002Fclosed\u002Fmerged) has not been\nre-verified in the same review is a live risk for a public post.\n\n## Severity and output\n\nGroup findings as **Blocking** (would make the announcement actively\nmisleading -- typically items 1, 2, and 3), **Should fix before announcing**\n(items 4, 5, 8), and **Worth tracking** (items 6, 7, if not independently\nblocking). For each finding: cite the file\u002Fsection, state the concrete\nscenario in which a reader or user would notice the gap, and give one\nspecific, actionable fix (soften a claim, add a caveat, write a missing\ntest, revise an estimate) rather than a vague \"consider clarifying.\"\n\nDo not treat \"the test suite passes\" or \"the doc-quality skill found\nnothing\" as evidence a control is ready to announce -- both are necessary,\nneither is sufficient, and this skill exists specifically to check what\nthey don't.\n\n## Troubleshooting\n\n| Symptom | Check |\n|---|---|\n| Every finding feels the same as `markdown-doc-quality`'s | You are re-checking prose quality, not the underlying claim. Ask \"has this actually been observed working,\" not \"is this well-written.\" |\n| No findings at all | Confirm you checked the control's actual authoritative signal against real, live-observed output, not against its test suite or design intent. A clean unit-test suite is not evidence here. |\n| Unsure whether a finding is Blocking | Ask: would a technical reader who tries the demo within an hour of the announcement notice this gap themselves? If yes, it is Blocking. |\n\n## Contributing\n\nChecklist items belong in this file and must stay portable across\ncontrols -- describe patterns by their behavior, not by a specific control\nID. Add a new item only after finding it live during an actual review, and\nname the control and date in the commit\u002FPR description, not in this file\nitself (the checklist stays generic; specific findings belong in the\ncontrol's own docs).\n","# Control Demo Roast Skill\n\n## Overview\n\nA control can pass every automated test, satisfy `markdown-doc-quality`'s\nchecklist, and still not deserve a public announcement, because none of\nthat checks whether the thing being announced actually does what the\nannouncement will claim. This skill is a deliberately adversarial, skeptical\npass over one control's full state (README, ASSESSMENT, code, tests, and\nwhat was actually run live) run **before** `linkedin-post-writer` drafts\nanything, and before a maintainer marks a control `Implemented`\u002F`Validated`\nin a change intended to be announced. Its job is to find the exact\nobjection a sharp member of the tech community would raise within the first\nfive minutes of actually trying the demo, and surface it now instead of in\npublic.\n\nThis complements `markdown-doc-quality` rather than replacing it:\n`markdown-doc-quality` catches writing-quality problems (hedging,\ncontradictions, missing evidence) in the prose itself. This skill catches a\ndifferent class of problem: prose that is well-written, internally\nconsistent, and passes every check, but describes a system whose central\nclaim has not actually been observed to hold. Apply both; a control can\nfail this skill's checklist while passing every item in that one.\n\n## When to use\n\n- Before drafting a `linkedin-post-writer` announcement for any control.\n- Before changing a control's `README.md` Status line to `Implemented` or\n  `Validated` in a change that will be publicized.\n- When a user asks to \"roast,\" \"sanity-check,\" or \"review before announcing\"\n  a control demo.\n- Not needed for a routine internal change that isn't headed toward a\n  public announcement or a status change.\n\n## The checklist\n\nWork through every item against the control's actual current state --\nREADME, ASSESSMENT.md, the real code, the real tests, and what was actually\nrun live this session -- not against what the control was designed to do.\nCite the exact file, section, or line each finding comes from; a finding\nwith no location is not actionable.\n\n#### 1. Status vs. observed reality\n\nDoes `Implemented`\u002F`Validated` describe what has actually been *observed\nworking*, or only what has been coded, unit-tested, and designed to work?\nFind the control's own authoritative signal (the thing its Control contract\nnames as \"Evidence \u002F source\") and ask: has this specific signal, from the\nspecific real service call the control depends on, ever actually been\nobserved producing its real output in a live run -- not a mocked, stubbed,\nor synthetic substitute? If the demo has only ever produced the *fail-closed*\nor *cannot-evaluate* branch of its own decision logic in every real run to\ndate, that is a specific, nameable gap: the control's core mechanism is\nunproven, not merely \"not yet exhaustively documented.\" State this\nplainly and do not let confident prose elsewhere in the README soften it.\n\n#### 2. Self-fulfilling or trivial proof\n\nWhen the demo's evidence includes a deliberately misconfigured or\nlower-quality component \"catching\" a problem, ask what was actually proven:\nthat the *monitoring\u002Fdetection mechanism* caught a real, independently-scored\nissue, or only that a component built to misbehave, in fact misbehaves. The\nsecond is not evidence the control's actual detection machinery works --\nit is evidence the demo's synthetic fixture is internally consistent. Name\nthis distinction explicitly when the \"breach\" scenario's detection came from\nconstruction (the reviewer already knows which component will fail) rather\nthan from the control's own authoritative signal actually flagging it\nindependently.\n\n#### 3. Unsubstantiated causal or outcome claims\n\nFlag any claim that a mechanism \"improves,\" \"raises,\" \"reduces,\" or \"heals\"\nsome real-world outcome over time, even when hedged (\"can help,\" \"may\nraise\"), when no experiment, measurement, or before\u002Fafter comparison in the\nsame control backs the causal claim. A plausible-sounding UX idea (for\nexample, showing a self-reported confidence score to a user) is not the\nsame as a demonstrated mechanism, and confident framing (\"a self-healing\nmechanism\") should not outrun what was actually measured. Recommend either\nsoftening the claim to what is actually true (a hypothesis, a design\nintent) or citing the actual evidence if it exists.\n\n#### 4. Complexity and time-estimate drift\n\nCompare the README's stated `Learning level` and `Estimated time` against\nthe *current* actual complexity: count real IAM role assignments, external\nresources, undocumented prerequisites, and any external dependency with\nunknown or open-ended latency (an async service that has never yet returned\nits result, RBAC propagation delays hit during the same session's live\ntesting, etc.). An estimate written when the design was simpler and never\nrevisited after it grew is a stale claim, not a current one. Recommend a\nrevised, honest estimate or an explicit caveat about the open-ended\ndependency.\n\n#### 5. Tone risk when citing an upstream gap\n\nWhen the control's writeup includes \"found a bug in the vendor SDK\" or \"the\nvendor's documentation has a gap,\" check whether the overall narrative could\nread as blaming the platform for the demo's own core mechanism not working,\nespecially when that mechanism is still unproven (see item 1). A real,\nindependently-reproduced upstream finding is worth citing professionally and\nfactually; it reads very differently next to an honest \"and our own\ncontrol doesn't fully work yet either\" than next to confident\n`Implemented` framing that omits it. Recommend adjusting emphasis, not\nremoving a legitimate finding.\n\n#### 6. Test coverage vs. what a first-time user will actually witness\n\nCheck whether any automated test pins down the *specific outcome* a person\nrunning the documented demo commands today will actually see -- not just\nclean unit tests of internal functions with synthetic inputs. If the real,\ncurrent, only-ever-observed outcome of running the demo is a specific\ndecision (for example, always `cannot_evaluate`), and no test asserts that\noutcome against the real integration point (not a mocked one), flag this as\na coverage gap distinct from \"tests exist and pass\": the tests do not\nprotect the one behavior most likely to surprise a new user.\n\n#### 7. Borrowed assumptions never confirmed against the real integration\n\nFlag any data shape, schema, or contract \"confirmed\" against a *different*\nsystem, mode, or code path (a legacy mechanism, a different API, an earlier\ndesign) and then assumed, without confirmation, to hold for the *new*\nmechanism the control now claims to use. If the new integration's real\noutput has never actually been observed, say so explicitly rather than\nletting a schema borrowed from elsewhere read as verified.\n\n#### 8. Stale or unverified external citations\n\nFor every external link or issue reference presented as current evidence\n(a GitHub issue, a vendor doc), confirm it was actually checked recently\nenough to still be accurate -- not simply carried forward from when it was\nfirst found. A citation whose status (open\u002Fclosed\u002Fmerged) has not been\nre-verified in the same review is a live risk for a public post.\n\n## Severity and output\n\nGroup findings as **Blocking** (would make the announcement actively\nmisleading -- typically items 1, 2, and 3), **Should fix before announcing**\n(items 4, 5, 8), and **Worth tracking** (items 6, 7, if not independently\nblocking). For each finding: cite the file\u002Fsection, state the concrete\nscenario in which a reader or user would notice the gap, and give one\nspecific, actionable fix (soften a claim, add a caveat, write a missing\ntest, revise an estimate) rather than a vague \"consider clarifying.\"\n\nDo not treat \"the test suite passes\" or \"the doc-quality skill found\nnothing\" as evidence a control is ready to announce -- both are necessary,\nneither is sufficient, and this skill exists specifically to check what\nthey don't.\n\n## Troubleshooting\n\n| Symptom | Check |\n|---|---|\n| Every finding feels the same as `markdown-doc-quality`'s | You are re-checking prose quality, not the underlying claim. Ask \"has this actually been observed working,\" not \"is this well-written.\" |\n| No findings at all | Confirm you checked the control's actual authoritative signal against real, live-observed output, not against its test suite or design intent. A clean unit-test suite is not evidence here. |\n| Unsure whether a finding is Blocking | Ask: would a technical reader who tries the demo within an hour of the announcement notice this gap themselves? If yes, it is Blocking. |\n\n## Contributing\n\nChecklist items belong in this file and must stay portable across\ncontrols -- describe patterns by their behavior, not by a specific control\nID. Add a new item only after finding it live during an actual review, and\nname the control and date in the commit\u002FPR description, not in this file\nitself (the checklist stays generic; specific findings belong in the\ncontrol's own docs).\n",{"name":5,"license":15,"metadata":24,"description":7,"user-invocable":27},{"authors":12,"last_updated":25,"spec_version":26},"2026-09-25","1.0",true,"main","e51c6b933afa0091b74eb2ae277466a8a39a929d",".claude\u002Fskills\u002Fcontrol-demo-roast",1,"https:\u002F\u002Fgithub.com\u002Fdoruit\u002Fforged-with-foundry","https:\u002F\u002Fgithub.com\u002Fdoruit","MVP Microsoft AI Platform","2026-11-15 00:34:55",{"source":8,"skillName":5,"commands":37,"prompt":47,"zipAvailable":27},[38,41,44],{"tool":39,"command":40},"npx","npx skills add doruit\u002Fforged-with-foundry --skill control-demo-roast",{"tool":42,"command":43},"bunx","bunx skills add doruit\u002Fforged-with-foundry --skill control-demo-roast",{"tool":45,"command":46},"pnpm","pnpm dlx skills add doruit\u002Fforged-with-foundry --skill control-demo-roast","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fdoruit\u002Fforged-with-foundry\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fcontrol-demo-roast\u002FSKILL.md",[],[50,57,64,71,78,83],{"skillNo":51,"name":52,"nameZh":6,"description":53,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":54,"categories":55,"keywords":56,"publishedAt":19},"01M7BV6R1JZ71GKYFK224G3343","linkedin-post-writer","Generates rich, engagement-optimized LinkedIn posts announcing new AI governance control demos, weekly\u002Fmonthly repo updates, and category-completion milestones for Forged with Foundry","https:\u002F\u002Fgithub.com\u002Fdoruit\u002Fforged-with-foundry\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Flinkedin-post-writer\u002FSKILL.md",[],[],{"skillNo":58,"name":59,"nameZh":6,"description":60,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":61,"categories":62,"keywords":63,"publishedAt":19},"01M7BV6M4TDHDYMYZSZZMMDC9K","markdown-doc-quality","Detects AI-slop patterns in Markdown documentation: heading\u002Fbody contradictions, unverified external claims, hedging filler, missing evidence examples, and inconsistent terminology across .md files","https:\u002F\u002Fgithub.com\u002Fdoruit\u002Fforged-with-foundry\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fcoding-standards\u002Fmarkdown-doc-quality\u002FSKILL.md",[],[],{"skillNo":65,"name":66,"nameZh":6,"description":67,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":68,"categories":69,"keywords":70,"publishedAt":19},"01M7BV6J19CM0HVX2RQN98R920","assess-governance-control","Assess a governance control before implementation","https:\u002F\u002Fgithub.com\u002Fdoruit\u002Fforged-with-foundry\u002Fblob\u002Fmain\u002F.github\u002Fskills\u002Fassess-governance-control\u002FSKILL.md",[],[],{"skillNo":72,"name":73,"nameZh":6,"description":74,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":19},"01M7BV68TFWNHFXA4T5XY5ZB08","mermaid-infra-diagram","Generate a control's 'Demo infrastructure setup (simplified)' diagram with real Microsoft\u002FAzure service icons instead of abstract colored boxes. Primary path: draw.io XML rendered headlessly via the bundled VS Code draw.io extension, using the official Azure Public Service Icons set indexed at media\u002Ficons\u002FAzure_Public_Service_Icons\u002Ficon-index.jsonl. Fallback path: Mermaid architecture-beta with Iconify icon packs, pre-rendered because GitHub does not render Mermaid icon packs live. For controls that deploy Azure resources, follow the KNW-001 reference layout: nested subscription, resource group and resource boundaries, a legend, an outside-scope band and tiered icon sizes. Use when creating or refreshing a control's Demo infrastructure setup (simplified) section, or when a plain flowchart diagram should show recognizable real-world service icons instead of abstract colored boxes.","https:\u002F\u002Fgithub.com\u002Fdoruit\u002Fforged-with-foundry\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fmermaid-infra-diagram\u002FSKILL.md",[],[],{"skillNo":79,"name":59,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":19},"01M7BV66RP0PDAED33QNKPFK2W","https:\u002F\u002Fgithub.com\u002Fdoruit\u002Fforged-with-foundry\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fmarkdown-doc-quality\u002FSKILL.md",[],[],{"skillNo":84,"name":52,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":85,"categories":86,"keywords":87,"publishedAt":19},"01M7BV64PGDQFN5K5YVGFV70ZG","https:\u002F\u002Fgithub.com\u002Fdoruit\u002Fforged-with-foundry\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Flinkedin-post-writer\u002FSKILL.md",[],[]]