[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7DEF7P82EMCVX2YSZR0ZGZQ":3,"skill-same:01M7DEF7P82EMCVX2YSZR0ZGZQ":56},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":14,"repositoryHtmlUrl":27,"repositoryHomepageUrl":28,"ownerHtmlUrl":29,"ownerBlogUrl":28,"ownerDescription":30,"updatedAt":31,"install":32,"resources":45},"01M7DEF7P82EMCVX2YSZR0ZGZQ","szl-governed-decision",null,"Make decisions that carry their own evidence. Wrap any classifier, policy engine or System One model in a fail-closed record that names why it refused, cites a digest rather than the text, and never asserts a number without a receipt on disk. Use when a decision must survive audit, when a model's output should be advisory rather than authoritative, or when a claim needs a state - MEASURED, BLOCKED, UNVERIFIED, UNAVAILABLE, NOT_CLAIMED - instead of a confident sentence. Complements typed-primitive skills: they produce the judgement, this governs what the judgement is allowed to do.","szl-holdings\u002Fszl-skills","szl-holdings","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F273360816?v=4","01M7DEC83S24HX3V8G1JSWKHD0","SZL Holdings","01M7DEC85084EGHEJH2KMAA53K",0,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fszl-holdings\u002Fszl-skills\u002Fblob\u002Fmain\u002Fskills\u002Fszl-governed-decision\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: szl-governed-decision\ndescription: \"Make decisions that carry their own evidence. Wrap any classifier, policy engine or System One model in a fail-closed record that names why it refused, cites a digest rather than the text, and never asserts a number without a receipt on disk. Use when a decision must survive audit, when a model's output should be advisory rather than authoritative, or when a claim needs a state - MEASURED, BLOCKED, UNVERIFIED, UNAVAILABLE, NOT_CLAIMED - instead of a confident sentence. Complements typed-primitive skills: they produce the judgement, this governs what the judgement is allowed to do.\"\n\n---\n\n# Governed decision\n\nA decision is not a label. It is a record: what was decided, by which path, on what evidence,\nand what remains unproven.\n\n## The four refusal paths\n\nEvery decision reports exactly one.\n\n- `PRE_AGGREGATION` — a guard fired before any score existed. Sub-reason required:\n  `INJECTION_GUARD` (adversarial pattern matched a versioned policy list, the model was not\n  consulted), `META_CUE` (an enumerated handling cue), or `UNIDENTIFIED`.\n- `ZERO_PINNED` — scores were computed and one was exactly zero, so the weighted geometric\n  aggregate is zero. Weakest-link, not average.\n- `BELOW_THRESHOLD` — all scores positive, aggregate under the floor.\n- `NONE` — a confident decision was issued.\n\n`UNIDENTIFIED` is not a fallback for convenience. If a refusal happens and no known guard\nexplains it, record `UNIDENTIFIED` and go find the guard. Naming a mechanism you have not\nlocated in the code is a guess wearing a label.\n\n## Never default a missing score\n\nAsking for a score vector on a decision that computed no scores must raise, not return zeros.\nIn a zero-pinned aggregate, a substituted zero is indistinguishable from a genuine veto — it\nmanufactures the exact condition the system treats as fatal.\n\n## The model provider contract\n\nA model attached to this kind of decision may:\n\n- lower one axis, never raise it — take the minimum, never the maximum;\n- justify a lowering with a span that occurs **literally** in the input, verified before the\n  score is honoured;\n- return nothing at all, in which case the axis is untouched and the reason is recorded.\n\nIt may not return a label. Under zero-pinning this makes a bad provider incapable of causing a\nconfident error: the worst it achieves is over-refusal, which shows up immediately as churn\nagainst a control corpus.\n\nConsult the provider only on decisions that reach the aggregator. A decision already refused by\na guard must never be credited to the provider.\n\n## Claim states\n\nAttach one to every assertion, and generate prose from the states rather than writing prose and\nhoping it matches.\n\n| state | meaning |\n|---|---|\n| MEASURED | a receipt exists on disk and the number was read from it |\n| UNRATIFIED | rests on labels no human has ratified |\n| DEGRADED | works, below the standard you hold yourself to |\n| BLOCKED | a negative result, recorded rather than omitted |\n| UNVERIFIED | unmeasured; a probability is a claim until calibration measures it |\n| UNAVAILABLE | no evidence exists; never rendered as health |\n| NOT_CLAIMED | a deliberate refusal to assert |\n\n## Rules that hold without exception\n\n1. Carry `sha256(input)`, never the input, into any receipt, span or telemetry attribute.\n2. Never fabricate a measurement you did not take. Energy is `null`, not `0`.\n3. An unsigned receipt says `signed: false`. It renders as UNAVAILABLE, never as a pass.\n4. Derive ids from content, not from clocks or randomness, so re-running produces no churn.\n5. Pin model versions. A moving `-latest` silently changes answers under a tuned threshold.\n6. Fix acceptance criteria before running the experiment, so they cannot be rewritten to match\n   the result.\n7. When a claim turns out wrong, add a correction with its receipt. Do not edit the old one\n   away.\n\n## Scripts\n\n- `scripts\u002Fdecide.py` — classify text, print the governed record and a receipt payload. It needs a szl-typesafe-triage checkout as its repository root.\n- `scripts\u002Fcheck_claims.py` — verify every claim in a ledger cites a file that exists.","# Governed decision\n\nA decision is not a label. It is a record: what was decided, by which path, on what evidence,\nand what remains unproven.\n\n## The four refusal paths\n\nEvery decision reports exactly one.\n\n- `PRE_AGGREGATION` — a guard fired before any score existed. Sub-reason required:\n  `INJECTION_GUARD` (adversarial pattern matched a versioned policy list, the model was not\n  consulted), `META_CUE` (an enumerated handling cue), or `UNIDENTIFIED`.\n- `ZERO_PINNED` — scores were computed and one was exactly zero, so the weighted geometric\n  aggregate is zero. Weakest-link, not average.\n- `BELOW_THRESHOLD` — all scores positive, aggregate under the floor.\n- `NONE` — a confident decision was issued.\n\n`UNIDENTIFIED` is not a fallback for convenience. If a refusal happens and no known guard\nexplains it, record `UNIDENTIFIED` and go find the guard. Naming a mechanism you have not\nlocated in the code is a guess wearing a label.\n\n## Never default a missing score\n\nAsking for a score vector on a decision that computed no scores must raise, not return zeros.\nIn a zero-pinned aggregate, a substituted zero is indistinguishable from a genuine veto — it\nmanufactures the exact condition the system treats as fatal.\n\n## The model provider contract\n\nA model attached to this kind of decision may:\n\n- lower one axis, never raise it — take the minimum, never the maximum;\n- justify a lowering with a span that occurs **literally** in the input, verified before the\n  score is honoured;\n- return nothing at all, in which case the axis is untouched and the reason is recorded.\n\nIt may not return a label. Under zero-pinning this makes a bad provider incapable of causing a\nconfident error: the worst it achieves is over-refusal, which shows up immediately as churn\nagainst a control corpus.\n\nConsult the provider only on decisions that reach the aggregator. A decision already refused by\na guard must never be credited to the provider.\n\n## Claim states\n\nAttach one to every assertion, and generate prose from the states rather than writing prose and\nhoping it matches.\n\n| state | meaning |\n|---|---|\n| MEASURED | a receipt exists on disk and the number was read from it |\n| UNRATIFIED | rests on labels no human has ratified |\n| DEGRADED | works, below the standard you hold yourself to |\n| BLOCKED | a negative result, recorded rather than omitted |\n| UNVERIFIED | unmeasured; a probability is a claim until calibration measures it |\n| UNAVAILABLE | no evidence exists; never rendered as health |\n| NOT_CLAIMED | a deliberate refusal to assert |\n\n## Rules that hold without exception\n\n1. Carry `sha256(input)`, never the input, into any receipt, span or telemetry attribute.\n2. Never fabricate a measurement you did not take. Energy is `null`, not `0`.\n3. An unsigned receipt says `signed: false`. It renders as UNAVAILABLE, never as a pass.\n4. Derive ids from content, not from clocks or randomness, so re-running produces no churn.\n5. Pin model versions. A moving `-latest` silently changes answers under a tuned threshold.\n6. Fix acceptance criteria before running the experiment, so they cannot be rewritten to match\n   the result.\n7. When a claim turns out wrong, add a correction with its receipt. Do not edit the old one\n   away.\n\n## Scripts\n\n- `scripts\u002Fdecide.py` — classify text, print the governed record and a receipt payload. It needs a szl-typesafe-triage checkout as its repository root.\n- `scripts\u002Fcheck_claims.py` — verify every claim in a ledger cites a file that exists.",{"name":5,"description":7},"main","8dd614e1f5097564340e36915f849d7d893cd68a","skills\u002Fszl-governed-decision","https:\u002F\u002Fgithub.com\u002Fszl-holdings\u002Fszl-skills","https:\u002F\u002Fa-11-oy.com","https:\u002F\u002Fgithub.com\u002Fszl-holdings","Governed AI infrastructure for inspectable decisions: source, runtime state, receipts, and proof boundaries.","2026-11-15 03:52:34",{"source":8,"skillName":5,"commands":33,"prompt":43,"zipAvailable":44},[34,37,40],{"tool":35,"command":36},"npx","npx skills add szl-holdings\u002Fszl-skills --skill szl-governed-decision",{"tool":38,"command":39},"bunx","bunx skills add szl-holdings\u002Fszl-skills --skill szl-governed-decision",{"tool":41,"command":42},"pnpm","pnpm dlx skills add szl-holdings\u002Fszl-skills --skill szl-governed-decision","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fszl-holdings\u002Fszl-skills\u002Fblob\u002Fmain\u002Fskills\u002Fszl-governed-decision\u002FSKILL.md",true,[46,52],{"path":47,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":50,"sourceHtmlUrl":51},"scripts\u002Fcheck_claims.py","SCRIPT","脚本",688,"https:\u002F\u002Fgithub.com\u002Fszl-holdings\u002Fszl-skills\u002Fblob\u002Fmain\u002Fskills\u002Fszl-governed-decision\u002Fscripts\u002Fcheck_claims.py",{"path":53,"resourceType":48,"resourceTypeDesc":49,"sizeBytes":54,"sourceHtmlUrl":55},"scripts\u002Fdecide.py",1296,"https:\u002F\u002Fgithub.com\u002Fszl-holdings\u002Fszl-skills\u002Fblob\u002Fmain\u002Fskills\u002Fszl-governed-decision\u002Fscripts\u002Fdecide.py",[57,64,71,78,84,89],{"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},"01M7DENG22R761RTDSG11WWJB3","szl-unit-invariants","Checks declared SI dimensions, allowlisted unit conversions, expected ranges and fixed pointwise invariants (equal, sum, difference, product, ratio) over selected measurement or simulation columns, flagging incompatible quantities, scale errors of 1000x and broken conservation checks. Use when merging instrument exports with different units, validating simulation output against conservation laws, or when a reviewer asks how unit errors were ruled out. Not an uncertainty-propagation or symbolic tool.","https:\u002F\u002Fgithub.com\u002Fszl-holdings\u002Fszl-skills\u002Fblob\u002Fmain\u002Fskills\u002Fszl-unit-invariants\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},"01M7DENAVBJ6GSCRT3E1PWK56D","szl-uncertainty-lineage","Propagate declared measurement uncertainty through a small arithmetic dependency graph using a first-order delta approximation. Use when a researcher needs to see how corrected inputs or shared-source correlation change a reported standard uncertainty; require unit checks separately.","https:\u002F\u002Fgithub.com\u002Fszl-holdings\u002Fszl-skills\u002Fblob\u002Fmain\u002Fskills\u002Fszl-uncertainty-lineage\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},"01M7DEN2M7P59RTM6RJE99PGN8","szl-typesafe-ai","TypeSafe\u002FJev: Choice\u002FNoul\u002FScore in code, not TypeScript types. System One models turn natural language and application state into probabilities that software can compose. Use when building routing, ranking, extraction, verification, or replacing an LLM prompt-and-parse step with structured decisions; also when brainstorming what TypeSafe could make possible. Do not use for TypeScript\u002FJava\u002FRust static types, Zod\u002FPydantic\u002Fmypy\u002FJSON Schema\u002Fprotobuf codegen, or asking whether a pipeline is type-safe. At SZL, Jev is an optional second reader only.","https:\u002F\u002Fgithub.com\u002Fszl-holdings\u002Fszl-skills\u002Fblob\u002Fmain\u002Fskills\u002Fszl-typesafe-ai\u002FSKILL.md",[],[],{"skillNo":79,"name":80,"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":81,"categories":82,"keywords":83,"publishedAt":19},"01M7DEN1C9G4H12N9HRHYK2HZ7","demo","https:\u002F\u002Fgithub.com\u002Fszl-holdings\u002Fszl-skills\u002Fblob\u002Fmain\u002Fskills\u002Fszl-skill-update-review\u002Fassets\u002Fexample\u002Fold-package\u002Fskills\u002Fdemo\u002FSKILL.md",[],[],{"skillNo":85,"name":80,"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":86,"categories":87,"keywords":88,"publishedAt":19},"01M7DEMZYRXT5CG2XPXERAM218","https:\u002F\u002Fgithub.com\u002Fszl-holdings\u002Fszl-skills\u002Fblob\u002Fmain\u002Fskills\u002Fszl-skill-update-review\u002Fassets\u002Fexample\u002Fnew-package\u002Fskills\u002Fdemo\u002FSKILL.md",[],[],{"skillNo":90,"name":91,"nameZh":6,"description":92,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":93,"categories":94,"keywords":95,"publishedAt":19},"01M7DEMVH2NMG4PS0DZTMGATAF","szl-skill-update-review","Compare pinned local skill packages using inventories and a retained lock; report changed skills, files, declarations and reruns without approving safety or validity.","https:\u002F\u002Fgithub.com\u002Fszl-holdings\u002Fszl-skills\u002Fblob\u002Fmain\u002Fskills\u002Fszl-skill-update-review\u002FSKILL.md",[],[]]