{"id":"67857c1f-7266-4f14-b82c-d7ff977f4fe5","shortId":"tebpv4","kind":"skill","title":"workbench-sdd","tagline":"Specification-driven development from raw requirement to product design, technical design, task list, execution, and verification.","description":"# Workbench SDD\n\nUse this skill when a task needs to move from a fuzzy request into executable work.\n\nThis is the default workbench planning protocol unless the issue explicitly says the task is a quick fix, emergency repair, or direct verification run.\n\n## Pipeline\n\nMove work through these stages:\n\n1. Raw Requirement\n   - Capture the user's literal request.\n   - Separate confirmed facts from assumptions.\n   - Name the owner, expected output, and known constraints.\n2. Product Design\n   - Define the user-facing behavior, success criteria, non-goals, and edge cases.\n   - Keep the scope narrow unless the user asks for expansion.\n3. Technical Design\n   - Identify the runtime owner, data path, files, commands, integrations, and risk surface.\n   - Prefer existing stable routes over new infrastructure.\n4. Task List\n   - Convert the design into bounded tasks with owner, evidence, and verification criteria.\n   - Keep tasks executable by one agent unless the issue explicitly asks for parallel work.\n   - Mark `GOAL_MODE: yes` when the owner must keep the objective alive across turns until verified.\n5. Execution And Verification\n   - Execute only the assigned slice.\n   - Verify on the real path.\n   - Report evidence before claiming done.\n\n## Stage Gate Rules\n\n- Do not jump from an ambiguous raw request directly to implementation.\n- Do not expand blast radius just because more agents or tokens are available.\n- If a stage is already answered by issue text or prior comments, cite that evidence and move forward.\n- If the task is blocked by a missing decision, state the smallest decision needed.\n- If the task is low-risk and obvious, compress the stages into a short SDD card instead of creating ceremony.\n- If the user says \"go\", \"continue\", or an equivalent approval after a reviewed stage, record that as the gate approval context before proceeding.\n- When a stage has both proxy and agent-authored artifacts, use the complete agent-authored artifact as primary and keep the proxy comment as trace evidence.\n- If an issue is marked `GOAL_MODE: yes`, the Task List must include `GOAL_LOCK`, closeout gates, and operator-call conditions before execution starts.\n\n## Comment Structure\n\nEach SDD stage is a structured issue comment, not an issue status. Use this header for every stage artifact:\n\n```text\nSDD_STAGE: [Raw Requirement / Product Design / Technical Design / Task List / Execution And Verification]\nOWNER: [one agent or human owner]\nSTATUS: READY_FOR_REVIEW / APPROVED / BLOCKED\nREVIEWER: Workbench Supervisor or designated reviewer\nEVIDENCE: [files, commands, issue/comment IDs, or artifacts checked]\nHANDOFF_SUMMARY: [five lines or fewer: what the next agent needs without rereading full history]\nSCOPED_EVIDENCE: [exact comment IDs, run IDs, commit hashes, files, or artifact paths to inspect]\nANTI_OVER_READ: [sources to skip unless needed, such as full issue lists or full comment history]\n```\n\nPut the stage-specific artifact after the header. Keep discussion replies separate from stage artifacts so the comment history remains scannable.\n\nEvery SDD stage comment must literally include the exact uppercase field names `HANDOFF_SUMMARY`, `SCOPED_EVIDENCE`, and `ANTI_OVER_READ` in the header. Do not substitute semantic equivalents such as \"Evidence\", \"Skipped\", \"Context\", or prose bullets. Every SDD review comment must literally include `VERDICT_SUMMARY`. The next agent should be able to start from the handoff summary alone and deep-read only the listed `SCOPED_EVIDENCE`.\n\n## Stage Gate Mechanics\n\n- Supervisor review happens between stages with `VERDICT: PASS / FLAG / BLOCK`.\n- `PASS` allows the next stage to start.\n- `FLAG` means the current stage needs a targeted correction before handoff.\n- `BLOCK` means the stage cannot proceed until the smallest blocking decision or missing input is resolved.\n- Issue statuses stay coarse: `todo` before work starts, `in_progress` while any SDD stage is active, `in_review` after execution is ready for final review, and `done` only after accepted evidence.\n\n## Bypass Rules\n\n- Workbench Admin may use `SDD_BYPASS: quick-fix` for low-risk, obvious changes where a full SDD sequence would add noise.\n- Workbench Admin may use `SDD_BYPASS: emergency` for time-critical repair work.\n- Bypass is not allowed for high-risk, ambiguous, multi-system, or public/private boundary changes.\n- A bypassed issue still needs one clear owner, explicit scope, verification evidence, and Supervisor review before closure.\n\n## Execution Handoff\n\n- The Task List stage names the execution owner, exact files/resources, non-goals, approval gates, and verification commands.\n- For `GOAL_MODE: yes`, the Task List also names required build/test/help-smoke/docs-or-report/git-status gates and the conditions for calling the operator.\n- Execution owners implement only their assigned slice and do not create follow-on work unless explicitly requested.\n- T8 or smoke-test work remains separate when the task list says it depends on a committed repo batch.\n- Completion reports include changed files, verification output, residual risks, commit hash or artifact link, work directory, and branch.\n- Live skill or agent-binding changes must have a backup and post-change verification before final PASS.\n- If a run reaches evidence-ready state but does not post a stage artifact, the conductor may post a proxy artifact using run-message evidence and mark it as proxy/recovery evidence.\n\n## Output Contract\n\nFor planning or routing, return:\n\n- `SDD_STAGE`: current stage.\n- `CONFIRMED`: facts and constraints.\n- `ASSUMPTIONS`: assumptions that still matter.\n- `NEXT_TASKS`: owner-scoped tasks.\n- `VERIFICATION`: evidence required to close.\n\nFor execution, return:\n\n- what changed,\n- what was verified,\n- what remains risky or blocked.","tags":["workbench","sdd","multica","ultimate","fearvox","agent-skills","ai-agents","claude-code","codex","hermes-agent","skills-sh"],"capabilities":["skill","source-fearvox","skill-workbench-sdd","topic-agent-skills","topic-ai-agents","topic-claude-code","topic-codex","topic-hermes-agent","topic-multica","topic-sdd","topic-skills-sh"],"categories":["multica-ultimate-workbench"],"synonyms":[],"warnings":[],"endpointUrl":"https://skills.sh/Fearvox/multica-ultimate-workbench/workbench-sdd","protocol":"skill","transport":"skills-sh","auth":{"type":"none","details":{"cli":"npx skills add Fearvox/multica-ultimate-workbench","source_repo":"https://github.com/Fearvox/multica-ultimate-workbench","install_from":"skills.sh"}},"qualityScore":"0.456","qualityRationale":"deterministic score 0.46 from registry signals: · indexed on github topic:agent-skills · 12 github stars · SKILL.md body (5,784 chars)","verified":false,"liveness":"unknown","lastLivenessCheck":null,"agentReviews":{"count":0,"score_avg":null,"cost_usd_avg":null,"success_rate":null,"latency_p50_ms":null,"narrative_summary":null,"summary_updated_at":null},"enrichmentModel":"deterministic:skill-github:v1","enrichmentVersion":1,"enrichedAt":"2026-05-18T19:07:14.823Z","embedding":null,"createdAt":"2026-05-13T01:05:04.228Z","updatedAt":"2026-05-18T19:07:14.823Z","lastSeenAt":"2026-05-18T19:07:14.823Z","tsv":"'1':69 '2':91 '3':118 '4':140 '5':185 'abl':540 'accept':633 'across':181 'activ':619 'add':658 'admin':638,661 'agent':160,226,315,322,397,430,537,805 'agent-author':314,321 'agent-bind':804 'aliv':180 'allow':571,676 'alon':547 'alreadi':235 'also':733 'ambigu':212,681 'answer':236 'anti':451,507 'approv':293,303,405,721 'artifact':317,324,380,419,447,473,483,795,834,841 'ask':115,165 'assign':192,750 'assumpt':82,868,869 'author':316,323 'avail':230 'backup':811 'batch':782 'behavior':99 'bind':806 'blast':221 'block':253,406,569,588,597,896 'bound':147 'boundari':687 'branch':800 'build/test/help-smoke/docs-or-report/git-status':736 'bullet':525 'bypass':635,642,665,673,690 'call':355,742 'cannot':592 'captur':72 'card':279 'case':107 'ceremoni':283 'chang':651,688,786,807,815,888 'check':420 'cite':243 'claim':202 'clear':695 'close':883 'closeout':350 'closur':705 'coars':607 'command':128,415,725 'comment':242,331,360,369,439,466,486,493,529 'commit':443,780,792 'complet':320,783 'compress':272 'condit':356,740 'conductor':836 'confirm':79,864 'constraint':90,867 'context':304,522 'continu':289 'contract':854 'convert':143 'correct':585 'creat':282,755 'criteria':101,154 'critic':670 'current':580,862 'data':125 'decis':257,261,598 'deep':550 'deep-read':549 'default':42 'defin':94 'depend':777 'design':13,15,93,120,145,387,389,411 'develop':7 'direct':60,215 'directori':798 'discuss':478 'done':203,630 'driven':6 'edg':106 'emerg':57,666 'equival':292,517 'everi':378,490,526 'evid':151,200,245,334,413,437,505,520,556,634,700,825,846,852,880 'evidence-readi':824 'exact':438,498,716 'execut':18,37,157,186,189,358,392,623,706,714,745,885 'exist':134 'expand':220 'expans':117 'expect':86 'explicit':49,164,697,761 'face':98 'fact':80,865 'fewer':426 'field':500 'file':127,414,445,787 'files/resources':717 'final':627,818 'five':423 'fix':56,645 'flag':568,577 'follow':757 'follow-on':756 'forward':248 'full':434,461,465,654 'fuzzi':34 'gate':205,302,351,558,722,737 'go':288 'goal':104,170,340,348,720,727 'handoff':421,502,545,587,707 'happen':562 'hash':444,793 'header':376,476,512 'high':679 'high-risk':678 'histori':435,467,487 'human':399 'id':417,440,442 'identifi':121 'implement':217,747 'includ':347,496,532,785 'infrastructur':139 'input':601 'inspect':450 'instead':280 'integr':129 'issu':48,163,238,337,368,372,462,604,691 'issue/comment':416 'jump':209 'keep':108,155,177,328,477 'known':89 'line':424 'link':796 'list':17,142,345,391,463,554,710,732,774 'liter':76,495,531 'live':801 'lock':349 'low':268,648 'low-risk':267,647 'mark':169,339,848 'matter':872 'may':639,662,837 'mean':578,589 'mechan':559 'messag':845 'miss':256,600 'mode':171,341,728 'move':31,64,247 'multi':683 'multi-system':682 'must':176,346,494,530,808 'name':83,501,712,734 'narrow':111 'need':29,262,431,458,582,693 'new':138 'next':429,536,573,873 'nois':659 'non':103,719 'non-goal':102,718 'object':179 'obvious':271,650 'one':159,396,694 'oper':354,744 'operator-cal':353 'output':87,789,853 'owner':85,124,150,175,395,400,696,715,746,876 'owner-scop':875 'parallel':167 'pass':567,570,819 'path':126,198,448 'pipelin':63 'plan':44,856 'post':814,831,838 'post-chang':813 'prefer':133 'primari':326 'prior':241 'proceed':306,593 'product':12,92,386 'progress':613 'prose':524 'protocol':45 'proxi':312,330,840 'proxy/recovery':851 'public/private':686 'put':468 'quick':55,644 'quick-fix':643 'radius':222 'raw':9,70,213,384 'reach':823 'read':453,509,551 'readi':402,625,826 'real':197 'record':298 'remain':488,769,893 'repair':58,671 'repli':479 'repo':781 'report':199,784 'request':35,77,214,762 'requir':10,71,385,735,881 'reread':433 'residu':790 'resolv':603 'return':859,886 'review':296,404,407,412,528,561,621,628,703 'risk':131,269,649,680,791 'riski':894 'rout':136,858 'rule':206,636 'run':62,441,822,844 'run-messag':843 'runtim':123 'say':50,287,775 'scannabl':489 'scope':110,436,504,555,698,877 'sdd':3,22,278,363,382,491,527,616,641,655,664,860 'semant':516 'separ':78,480,770 'sequenc':656 'short':277 'skill':25,802 'skill-workbench-sdd' 'skip':456,521 'slice':193,751 'smallest':260,596 'smoke':766 'smoke-test':765 'sourc':454 'source-fearvox' 'specif':5,472 'specification-driven':4 'stabl':135 'stage':68,204,233,274,297,309,364,379,383,471,482,492,557,564,574,581,591,617,711,833,861,863 'stage-specif':470 'start':359,542,576,611 'state':258,827 'status':373,401,605 'stay':606 'still':692,871 'structur':361,367 'substitut':515 'success':100 'summari':422,503,534,546 'supervisor':409,560,702 'surfac':132 'system':684 't8':763 'target':584 'task':16,28,52,141,148,156,251,265,344,390,709,731,773,874,878 'technic':14,119,388 'test':767 'text':239,381 'time':669 'time-crit':668 'todo':608 'token':228 'topic-agent-skills' 'topic-ai-agents' 'topic-claude-code' 'topic-codex' 'topic-hermes-agent' 'topic-multica' 'topic-sdd' 'topic-skills-sh' 'trace':333 'turn':182 'unless':46,112,161,457,760 'uppercas':499 'use':23,318,374,640,663,842 'user':74,97,114,286 'user-fac':96 'verdict':533,566 'verif':20,61,153,188,394,699,724,788,816,879 'verifi':184,194,891 'without':432 'work':38,65,168,610,672,759,768,797 'workbench':2,21,43,408,637,660 'workbench-sdd':1 'would':657 'yes':172,342,729","prices":[{"id":"5a910b9b-c7ca-4b6e-a9cb-7893fc0c0287","listingId":"67857c1f-7266-4f14-b82c-d7ff977f4fe5","amountUsd":"0","unit":"free","nativeCurrency":null,"nativeAmount":null,"chain":null,"payTo":null,"paymentMethod":"skill-free","isPrimary":true,"details":{"org":"Fearvox","category":"multica-ultimate-workbench","install_from":"skills.sh"},"createdAt":"2026-05-13T01:05:04.228Z"}],"sources":[{"listingId":"67857c1f-7266-4f14-b82c-d7ff977f4fe5","source":"github","sourceId":"Fearvox/multica-ultimate-workbench/workbench-sdd","sourceUrl":"https://github.com/Fearvox/multica-ultimate-workbench/tree/main/skills/workbench-sdd","isPrimary":false,"firstSeenAt":"2026-05-13T01:05:04.228Z","lastSeenAt":"2026-05-18T19:07:14.823Z"}],"details":{"listingId":"67857c1f-7266-4f14-b82c-d7ff977f4fe5","quickStartSnippet":null,"exampleRequest":null,"exampleResponse":null,"schema":null,"openapiUrl":null,"agentsTxtUrl":null,"citations":[],"useCases":[],"bestFor":[],"notFor":[],"kindDetails":{"org":"Fearvox","slug":"workbench-sdd","github":{"repo":"Fearvox/multica-ultimate-workbench","stars":12,"topics":["agent-skills","ai-agents","claude-code","codex","hermes-agent","multica","sdd","skills-sh"],"license":null,"html_url":"https://github.com/Fearvox/multica-ultimate-workbench","pushed_at":"2026-05-14T23:12:50Z","description":"Public multi-agent workbench skill pack for SDD, review gates, context discipline, VM lanes, Sanity context, and autonomous agent coordination.","skill_md_sha":"aa39d436d9745385e6d40dcc52e400d6740c559b","skill_md_path":"skills/workbench-sdd/SKILL.md","default_branch":"main","skill_tree_url":"https://github.com/Fearvox/multica-ultimate-workbench/tree/main/skills/workbench-sdd"},"layout":"multi","source":"github","category":"multica-ultimate-workbench","frontmatter":{"name":"workbench-sdd","description":"Specification-driven development from raw requirement to product design, technical design, task list, execution, and verification."},"skills_sh_url":"https://skills.sh/Fearvox/multica-ultimate-workbench/workbench-sdd"},"updatedAt":"2026-05-18T19:07:14.823Z"}}