[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7APTPEZJJQZJWWMND4DDPAN":3,"skill-same:01M7APTPEZJJQZJWWMND4DDPAN":51},{"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":34,"ownerDescription":35,"updatedAt":36,"install":37,"resources":50},"01M7APTPEZJJQZJWWMND4DDPAN","atlas-locate",null,"Phase L (Locate) — bounded probes and scatter subagents to answer DECISION_TARGET sub-questions. Descends a symbol → graph → lexical → windowed-read → test-dry-run ladder; emits FINDING-XXX records with evidence anchors and confidence tiers. Hot phase (≤60% of mission budget). Use after map.md is complete and sub-questions remain. Trigger phrases — \"find where X happens\", \"trace Y\", \"who writes to Z\", \"who calls this\", \"what implements this\".","Rynaro\u002Fmagicite","Rynaro","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F5271543?v=4","01M70CFYFMB316W23THVQAQ42M","Henrique A. Lavezzo","01M7APN2KCVWBNY727T57KR2ZJ",2,"Apache-2.0","https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002F.eidolons\u002Fatlas\u002Fskills\u002Flocate\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: atlas-locate\ndescription: Phase L (Locate) — bounded probes and scatter subagents to answer DECISION_TARGET sub-questions. Descends a symbol → graph → lexical → windowed-read → test-dry-run ladder; emits FINDING-XXX records with evidence anchors and confidence tiers. Hot phase (≤60% of mission budget). Use after map.md is complete and sub-questions remain. Trigger phrases — \"find where X happens\", \"trace Y\", \"who writes to Z\", \"who calls this\", \"what implements this\".\nallowed-tools: search_symbol graph_query search_text view_file test_dry_run memex_read\nmetadata:\n  methodology: ATLAS\n  phase: L\n---\n\n# SKILL: Locate — bounded probes & scatter subagents (Phase L)\n\n## When to use\n\n**Load when:** `map.md` is complete and sub-questions exist against `DECISION_TARGET`.\n\n**Unload when:** every sub-question has a `FINDING-XXX` at confidence ≥ M or is\nrecorded in `GAPS`.\n\n---\n\n## Contract\n\n| Field | Value |\n|-------|-------|\n| LLM calls permitted | Yes, but at the synthesis edge only. Retrieval remains deterministic-first. |\n| Tool budget | ≤ 60% of mission `max_tool_calls` — this is the hot phase |\n| Output | `findings.md` — list of `FINDING-XXX` records, each schema-valid |\n\n---\n\n## Probe ladder\n\nFor every sub-question, descend this ladder. Stop at the first tier that\nproduces a confident answer.\n\n### Tier 1 — Symbol probe (deterministic)\n\n```\nsearch_symbol(name=\"CastVoteRecord\")\n  → definition site\n  → all references (with path:line)\n  → subclasses \u002F implementers\n  → method definitions on the symbol\n```\n\nThis is a constant-cost index lookup. Use it whenever the sub-question\nreferences a named thing (class, method, constant, config key).\n\n### Tier 2 — Graph query (deterministic)\n\n```\ngraph_query(\"callers_of: RecordVote#call\")\ngraph_query(\"implementers_of: Tallier\")\ngraph_query(\"writers_to: cast_vote_records\")    # if graph indexes DB ops\n```\n\nGraph queries answer *relational* sub-questions without reading source.\n\n#### Graph-first decomposition (raise η before any windowed read)\n\nThe 90\u002F10 deterministic-first rule (I-3) means a relational sub-question should\nbe **exhausted on the graph** before any Tier-4 read. Codify the verb\nvocabulary and reach for it first:\n\n| Verb | Answers |\n|------|---------|\n| `callers_of(sym)` \u002F `callees_of(sym)` | direct call edges in\u002Fout of a symbol |\n| `implementers_of(iface)` \u002F `subclasses_of(class)` | polymorphic dispatch targets |\n| `writers_to(table)` \u002F `readers_of(table)` | data-flow endpoints (if graph indexes DB ops) |\n| `importers_of(module)` \u002F `imported_by(module)` | module-coupling edges |\n| `transitive_callers(sym, depth=N)` | depth-bounded upstream reach (cap the depth — unbounded = a read in disguise) |\n| `callgraph_slice(scope)` | the sub-graph induced by a scope glob — used to **partition** a surface |\n\nPushing more Locate work onto these deterministic structural probes is the\nsingle biggest lever on search-efficiency η: each is an O(1) index lookup with\nexact results, where an LLM-authored search is expensive and only\nprobabilistically correct.\n\n**Partition derives from the graph, not from a guess.** When the surface is\nlarge enough to scatter (`skills\u002Fscatter\u002FSKILL.md`), the per-module fan-out partition\nis computed from **one** parent-side `callgraph_slice(scope)` — the disjoint\nclusters are a structural fact. Do **not** LLM-guess the clustering; that\nre-introduces the inference trap the deterministic-first rule exists to avoid.\n\n### Tier 3 — Scoped lexical search\n\n```\nsearch_text(pattern=r'audit_log\\\\s*\\\\.\\\\s*write', scope='app\u002Fflows\u002F**', limit=50)\n```\n\nCaps at 50. If it returns 50, **do not** raise the cap. Two options:\n\n- Narrow the scope glob.\n- Replace the lexical pattern with a Tier 1\u002F2 probe.\n\nRepeated overflow = the question is mis-framed; revise.\n\n### Tier 4 — Windowed read\n\n```\nview_file(path='app\u002Fflows\u002Fvote_casting\u002Frecord_vote.rb', start=1, end=100)\n# if answer is beyond line 100, cursor forward\nview_file(path='...', start=100, end=200)\n```\n\nUse windowed reads only after tiers 1–3 have narrowed the target. Reading an\nentire file is a Phase-T mistake, not a probe.\n\n### Tier 5 — Dry-run or fixture inspection\n\nIf the mission concerns runtime behavior:\n\n```\ntest_dry_run(path='test\u002Fflows\u002Frecord_vote_test.rb', case='when user is guest')\n```\n\nTest files often document intended behavior more cleanly than implementation.\n\n---\n\n## The Operator pattern: scatter subagents\n\nWhen the mission has ≥2 independent sub-questions, do not pursue them\nserially in your own context. Spawn subagents.\n\n> For a **large surface** (> 5 modules OR > 25 files) with ≥2 disjoint\n> sub-questions, this scatter is formalized as the TRANCE-gated **Scatter-Gather\n> Locate** sub-mode — see `skills\u002Fscatter\u002FSKILL.md` for the both-flags activation\n> trigger, the 5-branch fan-out cap, the deterministic graph-derived partition,\n> and the merge+dedup contract. Below the threshold (or for a single \u002F\n> tightly-coupled sub-question), stay serial — the heuristics here apply.\n\n### When to scatter\n\n- Sub-questions are topologically disjoint (different modules, different\n  concerns).\n- Each sub-question is self-contained enough to answer with a small excerpt\n  of the map.\n- Total expected token cost > 15% of remaining budget if pursued serially.\n\n### Subagent spec\n\nEach subagent receives:\n\n```yaml\nsub_mission:\n  parent_mission_id: \u003CMISSION-ID>\n  question: \u003Cnatural-language sub-question>\n  scope: \u003Cpath globs>\n  budget:\n    max_tool_calls: \u003Cparent_budget \u002F num_subagents>\n    max_tokens_input: \u003Cparent_budget \u002F num_subagents>\n  required_finding_count: 1..N\n  map_excerpt: \u003Conly the MAP-MODULES and MAP-GRAPH slice relevant to scope>\n```\n\nEach subagent returns exactly one object:\n\n```yaml\nreturn:\n  findings: [FINDING-XXX, ...]\n  gaps: [...]\n  telemetry: {tokens_used, tool_calls}\n  # NO transcript. NO raw tool outputs.\n```\n\nThe parent context merges the `findings` list. Subagent transcripts are\n**destroyed**.\n\n### When NOT to scatter\n\n- Sub-question requires tight iteration with another sub-question.\n- Only one sub-question exists.\n- Budget is already tight; the coordination overhead isn't worth it.\n\n---\n\n## Writing a FINDING\n\n```yaml\nFINDING-017:\n  claim: \"RecordVote#call is the sole writer to cast_vote_records; authorization\n          is delegated to VotingAuthorizer#authorize_ballot.\"\n  evidence:\n    - path: app\u002Fflows\u002Fvote_casting\u002Frecord_vote.rb\n      lines: 42-78\n      excerpt_ref: memex:\u002F\u002Fexcerpt\u002F7f31ab\n    - path: app\u002Fpolicies\u002Fvoting_authorizer.rb\n      lines: 12-40\n      excerpt_ref: memex:\u002F\u002Fexcerpt\u002F2a9c05\n  ruled_out:\n    - \"No other CallNode writing to cast_vote_records found in app\u002F** (symbol\n       probe exhaustive).\"\n  confidence: H\n  supports_decision: DT-1\n```\n\n**Required fields:** `claim`, `evidence[]`, `confidence`, `supports_decision`.\n\n**`ruled_out` is first-class.** \"I searched X and found nothing\" is a finding\n— often the most valuable one. Do not discard negative results.\n\n**Excerpts go to Memex, not into findings.md.** The finding carries the\nreference; the raw text lives in the stable store.\n\n---\n\n## Confidence calibration\n\n| Tier | When |\n|------|------|\n| H | Claim is directly observed in an AST-derived structural fact or in explicit source text cited verbatim. |\n| M | Claim requires a short inferential step over observed evidence (e.g., \"this FlowObject *likely* runs in the request thread because it has no Sidekiq wrapping\"). |\n| L | Claim is plausible from context but not directly anchored. Flag in `GAPS`. |\n\nIf you catch yourself writing \"probably\", \"seems to\", \"I believe\" — it's not\nH. Be honest; downstream agents calibrate on your tiers.\n\n---\n\n## Three-strike halt\n\nFor each sub-question, maintain a strike counter. A probe returning\n`confidence: L` or empty results = one strike. Three strikes = halt the\nsub-question, record it in `GAPS` with:\n\n```yaml\nGAP-003:\n  sub_question: \"Does the voting FlowObject ever defer to Sidekiq?\"\n  strikes:\n    - probe: graph_query(\"callers_of: Sidekiq::Client.push in app\u002Fflows\u002Fvoting\u002F**\")\n      result: empty\n    - probe: search_text(\"perform_async|perform_later\", scope='app\u002Fflows\u002Fvoting\u002F**')\n      result: empty\n    - probe: view_file('app\u002Fflows\u002Fvoting\u002Frecord_vote.rb', 1, 100)\n      result: no Sidekiq references; synchronous only\n  working_hypothesis: \"Synchronous in the request thread.\"\n  escalate_to: human | follow-up-mission\n```\n\nThis is the dead-end guard. Without it, the agent will fixate.\n\n---\n\n## Exit gate\n\n- [ ] Every `DECISION_TARGET` sub-question has ≥1 finding at confidence ≥ M, OR\n      is recorded in `GAPS`.\n- [ ] Every finding has `path:line` anchors.\n- [ ] Every finding's `excerpt_ref` resolves in Memex.\n- [ ] No finding's claim contains hedging language without an L tier.\n- [ ] Tool-call count ≤ 60% of mission budget.\n","# SKILL: Locate — bounded probes & scatter subagents (Phase L)\n\n## When to use\n\n**Load when:** `map.md` is complete and sub-questions exist against `DECISION_TARGET`.\n\n**Unload when:** every sub-question has a `FINDING-XXX` at confidence ≥ M or is\nrecorded in `GAPS`.\n\n---\n\n## Contract\n\n| Field | Value |\n|-------|-------|\n| LLM calls permitted | Yes, but at the synthesis edge only. Retrieval remains deterministic-first. |\n| Tool budget | ≤ 60% of mission `max_tool_calls` — this is the hot phase |\n| Output | `findings.md` — list of `FINDING-XXX` records, each schema-valid |\n\n---\n\n## Probe ladder\n\nFor every sub-question, descend this ladder. Stop at the first tier that\nproduces a confident answer.\n\n### Tier 1 — Symbol probe (deterministic)\n\n```\nsearch_symbol(name=\"CastVoteRecord\")\n  → definition site\n  → all references (with path:line)\n  → subclasses \u002F implementers\n  → method definitions on the symbol\n```\n\nThis is a constant-cost index lookup. Use it whenever the sub-question\nreferences a named thing (class, method, constant, config key).\n\n### Tier 2 — Graph query (deterministic)\n\n```\ngraph_query(\"callers_of: RecordVote#call\")\ngraph_query(\"implementers_of: Tallier\")\ngraph_query(\"writers_to: cast_vote_records\")    # if graph indexes DB ops\n```\n\nGraph queries answer *relational* sub-questions without reading source.\n\n#### Graph-first decomposition (raise η before any windowed read)\n\nThe 90\u002F10 deterministic-first rule (I-3) means a relational sub-question should\nbe **exhausted on the graph** before any Tier-4 read. Codify the verb\nvocabulary and reach for it first:\n\n| Verb | Answers |\n|------|---------|\n| `callers_of(sym)` \u002F `callees_of(sym)` | direct call edges in\u002Fout of a symbol |\n| `implementers_of(iface)` \u002F `subclasses_of(class)` | polymorphic dispatch targets |\n| `writers_to(table)` \u002F `readers_of(table)` | data-flow endpoints (if graph indexes DB ops) |\n| `importers_of(module)` \u002F `imported_by(module)` | module-coupling edges |\n| `transitive_callers(sym, depth=N)` | depth-bounded upstream reach (cap the depth — unbounded = a read in disguise) |\n| `callgraph_slice(scope)` | the sub-graph induced by a scope glob — used to **partition** a surface |\n\nPushing more Locate work onto these deterministic structural probes is the\nsingle biggest lever on search-efficiency η: each is an O(1) index lookup with\nexact results, where an LLM-authored search is expensive and only\nprobabilistically correct.\n\n**Partition derives from the graph, not from a guess.** When the surface is\nlarge enough to scatter (`skills\u002Fscatter\u002FSKILL.md`), the per-module fan-out partition\nis computed from **one** parent-side `callgraph_slice(scope)` — the disjoint\nclusters are a structural fact. Do **not** LLM-guess the clustering; that\nre-introduces the inference trap the deterministic-first rule exists to avoid.\n\n### Tier 3 — Scoped lexical search\n\n```\nsearch_text(pattern=r'audit_log\\\\s*\\\\.\\\\s*write', scope='app\u002Fflows\u002F**', limit=50)\n```\n\nCaps at 50. If it returns 50, **do not** raise the cap. Two options:\n\n- Narrow the scope glob.\n- Replace the lexical pattern with a Tier 1\u002F2 probe.\n\nRepeated overflow = the question is mis-framed; revise.\n\n### Tier 4 — Windowed read\n\n```\nview_file(path='app\u002Fflows\u002Fvote_casting\u002Frecord_vote.rb', start=1, end=100)\n# if answer is beyond line 100, cursor forward\nview_file(path='...', start=100, end=200)\n```\n\nUse windowed reads only after tiers 1–3 have narrowed the target. Reading an\nentire file is a Phase-T mistake, not a probe.\n\n### Tier 5 — Dry-run or fixture inspection\n\nIf the mission concerns runtime behavior:\n\n```\ntest_dry_run(path='test\u002Fflows\u002Frecord_vote_test.rb', case='when user is guest')\n```\n\nTest files often document intended behavior more cleanly than implementation.\n\n---\n\n## The Operator pattern: scatter subagents\n\nWhen the mission has ≥2 independent sub-questions, do not pursue them\nserially in your own context. Spawn subagents.\n\n> For a **large surface** (> 5 modules OR > 25 files) with ≥2 disjoint\n> sub-questions, this scatter is formalized as the TRANCE-gated **Scatter-Gather\n> Locate** sub-mode — see `skills\u002Fscatter\u002FSKILL.md` for the both-flags activation\n> trigger, the 5-branch fan-out cap, the deterministic graph-derived partition,\n> and the merge+dedup contract. Below the threshold (or for a single \u002F\n> tightly-coupled sub-question), stay serial — the heuristics here apply.\n\n### When to scatter\n\n- Sub-questions are topologically disjoint (different modules, different\n  concerns).\n- Each sub-question is self-contained enough to answer with a small excerpt\n  of the map.\n- Total expected token cost > 15% of remaining budget if pursued serially.\n\n### Subagent spec\n\nEach subagent receives:\n\n```yaml\nsub_mission:\n  parent_mission_id: \u003CMISSION-ID>\n  question: \u003Cnatural-language sub-question>\n  scope: \u003Cpath globs>\n  budget:\n    max_tool_calls: \u003Cparent_budget \u002F num_subagents>\n    max_tokens_input: \u003Cparent_budget \u002F num_subagents>\n  required_finding_count: 1..N\n  map_excerpt: \u003Conly the MAP-MODULES and MAP-GRAPH slice relevant to scope>\n```\n\nEach subagent returns exactly one object:\n\n```yaml\nreturn:\n  findings: [FINDING-XXX, ...]\n  gaps: [...]\n  telemetry: {tokens_used, tool_calls}\n  # NO transcript. NO raw tool outputs.\n```\n\nThe parent context merges the `findings` list. Subagent transcripts are\n**destroyed**.\n\n### When NOT to scatter\n\n- Sub-question requires tight iteration with another sub-question.\n- Only one sub-question exists.\n- Budget is already tight; the coordination overhead isn't worth it.\n\n---\n\n## Writing a FINDING\n\n```yaml\nFINDING-017:\n  claim: \"RecordVote#call is the sole writer to cast_vote_records; authorization\n          is delegated to VotingAuthorizer#authorize_ballot.\"\n  evidence:\n    - path: app\u002Fflows\u002Fvote_casting\u002Frecord_vote.rb\n      lines: 42-78\n      excerpt_ref: memex:\u002F\u002Fexcerpt\u002F7f31ab\n    - path: app\u002Fpolicies\u002Fvoting_authorizer.rb\n      lines: 12-40\n      excerpt_ref: memex:\u002F\u002Fexcerpt\u002F2a9c05\n  ruled_out:\n    - \"No other CallNode writing to cast_vote_records found in app\u002F** (symbol\n       probe exhaustive).\"\n  confidence: H\n  supports_decision: DT-1\n```\n\n**Required fields:** `claim`, `evidence[]`, `confidence`, `supports_decision`.\n\n**`ruled_out` is first-class.** \"I searched X and found nothing\" is a finding\n— often the most valuable one. Do not discard negative results.\n\n**Excerpts go to Memex, not into findings.md.** The finding carries the\nreference; the raw text lives in the stable store.\n\n---\n\n## Confidence calibration\n\n| Tier | When |\n|------|------|\n| H | Claim is directly observed in an AST-derived structural fact or in explicit source text cited verbatim. |\n| M | Claim requires a short inferential step over observed evidence (e.g., \"this FlowObject *likely* runs in the request thread because it has no Sidekiq wrapping\"). |\n| L | Claim is plausible from context but not directly anchored. Flag in `GAPS`. |\n\nIf you catch yourself writing \"probably\", \"seems to\", \"I believe\" — it's not\nH. Be honest; downstream agents calibrate on your tiers.\n\n---\n\n## Three-strike halt\n\nFor each sub-question, maintain a strike counter. A probe returning\n`confidence: L` or empty results = one strike. Three strikes = halt the\nsub-question, record it in `GAPS` with:\n\n```yaml\nGAP-003:\n  sub_question: \"Does the voting FlowObject ever defer to Sidekiq?\"\n  strikes:\n    - probe: graph_query(\"callers_of: Sidekiq::Client.push in app\u002Fflows\u002Fvoting\u002F**\")\n      result: empty\n    - probe: search_text(\"perform_async|perform_later\", scope='app\u002Fflows\u002Fvoting\u002F**')\n      result: empty\n    - probe: view_file('app\u002Fflows\u002Fvoting\u002Frecord_vote.rb', 1, 100)\n      result: no Sidekiq references; synchronous only\n  working_hypothesis: \"Synchronous in the request thread.\"\n  escalate_to: human | follow-up-mission\n```\n\nThis is the dead-end guard. Without it, the agent will fixate.\n\n---\n\n## Exit gate\n\n- [ ] Every `DECISION_TARGET` sub-question has ≥1 finding at confidence ≥ M, OR\n      is recorded in `GAPS`.\n- [ ] Every finding has `path:line` anchors.\n- [ ] Every finding's `excerpt_ref` resolves in Memex.\n- [ ] No finding's claim contains hedging language without an L tier.\n- [ ] Tool-call count ≤ 60% of mission budget.\n",{"name":5,"metadata":24,"description":7,"allowed-tools":27},{"phase":25,"methodology":26},"L","ATLAS","search_symbol graph_query search_text view_file test_dry_run memex_read","main","38c3a82583abf9806de764f89ac0099ad690933a",".eidolons\u002Fatlas\u002Fskills\u002Flocate",0,"https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite","https:\u002F\u002Fgithub.com\u002FRynaro","https:\u002F\u002Fhlavezzo.me","A Code Alchemist with a passion for weaving tales and inscribing transmutation circles in pure code.","2026-10-14 02:35:37",{"source":8,"skillName":5,"commands":38,"prompt":48,"zipAvailable":49},[39,42,45],{"tool":40,"command":41},"npx","npx skills add Rynaro\u002Fmagicite --skill atlas-locate",{"tool":43,"command":44},"bunx","bunx skills add Rynaro\u002Fmagicite --skill atlas-locate",{"tool":46,"command":47},"pnpm","pnpm dlx skills add Rynaro\u002Fmagicite --skill atlas-locate","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002F.eidolons\u002Fatlas\u002Fskills\u002Flocate\u002FSKILL.md",true,[],[52,59,66,73,80,87],{"skillNo":53,"name":54,"nameZh":6,"description":55,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":56,"categories":57,"keywords":58,"publishedAt":19},"01M7AQ4GHJQVV8TJBJ8T1V7ZQ9","wine-dxvk-cache-clear","Clear a stale DXVK shader cache that is causing graphical corruption in a\nWine or Proton game. Use when: a game shows corrupted textures or shader\nflicker that started after a driver or DXVK version change.","https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Ftoy-registry\u002Fskills\u002Fwine-dxvk-cache-clear\u002FSKILL.md",[],[],{"skillNo":60,"name":61,"nameZh":6,"description":62,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":63,"categories":64,"keywords":65,"publishedAt":19},"01M7AQ4CY1NHSMY1HCZ91387YY","steam-download-region-fix","Fix a Steam client stuck at very low download speed by changing its\ndownload region. Use when: Steam downloads are far below the local\ninternet speed for every game, not just one.","https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Ftoy-registry\u002Fskills\u002Fsteam-download-region-fix\u002FSKILL.md",[],[],{"skillNo":67,"name":68,"nameZh":6,"description":69,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":19},"01M7AQ49P7Y0CDRSYMPVABX80K","proton-battleye-eac-toggle","Toggle BattlEye or Easy Anti-Cheat runtime support for a Steam Proton game.\nUse when: a multiplayer game refuses to launch with an anti-cheat error under Proton.","https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002Ftests\u002Ffixtures\u002Ftoy-registry\u002Fskills\u002Fproton-battleye-eac-toggle\u002FSKILL.md",[],[],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":19},"01M7AQ45326FAP3X4Q86FK1X4N","vivi-verify-incoming","Load when reading any upstream artefact handed off to Vivi that carries a sibling .envelope.json. BLOCKING per ECL §6.2.2: the orchestrator MUST have verified the envelope's SHA-256 (eidolons verify-envelope --block \u002F eidolons run --verify) and recorded a verify_pass before dispatch. If no verify_pass exists for the message_id, or a verify_fail is present, REFUSE to process the payload and hand back to the orchestrator. Symmetric receiver gate — every Eidolon enforces it identically.","https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002F.eidolons\u002Fvivi\u002Fskills\u002Fverify-incoming\u002FSKILL.md",[],[],{"skillNo":81,"name":82,"nameZh":6,"description":83,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":84,"categories":85,"keywords":86,"publishedAt":19},"01M7AQ402861P0P3C2TEGFGMA1","vivi-parallel-tracks","Load ONLY under TRANCE authorization when the Plan phase yields N independent, non-overlapping implementation tracks (disjoint file sets). Operationalizes the TRANCE G4 form — parallel multi-track implementation in mandatory git-worktree isolation, max 5 clean-context tracks, a per-track verifier cascade, a non-fungible per-track ≤3 reflection budget, explicit stop conditions, and a single-threaded merge\u002Faggregation step. Absent TRANCE gating, Vivi runs exactly as today: single-track A→P→I→V→Δ\u002FR.","https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002F.eidolons\u002Fvivi\u002Fskills\u002Fparallel-tracks\u002FSKILL.md",[],[],{"skillNo":88,"name":89,"nameZh":6,"description":90,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":91,"categories":92,"keywords":93,"publishedAt":19},"01M7AQ3RH9BZF9ZMTC7VG5P7PB","vivi-methodology","Full Vivi cycle reference — A(nalyze) → P(lan) → I(mplement) → V(erify) → Δ(Delta)\u002FR(eflect). Use when the task is a non-trivial feature implementation in a brownfield codebase and you need the complete methodology reference: complexity routing, evidence-grounded planning rules, test-anchoring requirements, and failure-escalation thresholds.","https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002F.eidolons\u002Fvivi\u002Fskills\u002Fmethodology\u002FSKILL.md",[],[]]