[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7APYP0C8FWB3B4STDF96NTK":3,"skill-same:01M7APYP0C8FWB3B4STDF96NTK":50},{"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":27,"commitSha":28,"skillDirectory":29,"forksCount":30,"repositoryHtmlUrl":31,"repositoryHomepageUrl":6,"ownerHtmlUrl":32,"ownerBlogUrl":33,"ownerDescription":34,"updatedAt":35,"install":36,"resources":49},"01M7APYP0C8FWB3B4STDF96NTK","kupo-patch-verify",null,"Phase P+O sandbox patch loop with external verifier; governs edit emission, harness applier invocation, per-file loop detector, circuit-breaker, and pre-completion green-signal gate. Use when Kupo enters Phase P to emit a patch or Phase O to run verifiers; load after a KEEP decision. Do NOT use during Phase K or U.","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\u002Fkupo\u002Fskills\u002Fpatch-verify\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: kupo-patch-verify\ndescription: Phase P+O sandbox patch loop with external verifier; governs edit emission, harness applier invocation, per-file loop detector, circuit-breaker, and pre-completion green-signal gate. Use when Kupo enters Phase P to emit a patch or Phase O to run verifiers; load after a KEEP decision. Do NOT use during Phase K or U.\nmetadata:\n  methodology: Kupo\n  phase: P+O\n---\n\n# Patch-Verify Skill — Kupo (Phase P+O loop)\n\n## When to use\n\nLoad during Phase P (Patch) and Phase O (Observe). Governs the\nedit → harness-apply → sandbox → verify loop. Never invoked in K or U.\n\nLoaded during Phase P (Patch) and Phase O (Observe). Governs the\nedit → harness-apply → sandbox → verify loop. Never invoked in K or U.\n\n---\n\n## Phase P — Patch\n\n### Edit emission\n\nEmit the edit in one of two forms. Never emit a diff.\n\n**Form 1 — search\u002Freplace (default):**\n\n```json\n{\n  \"target_path\": \"src\u002Ffoo.ts\",\n  \"edit_kind\": \"search_replace\",\n  \"blocks\": [\n    { \"search\": \"\u003Cexact verbatim text to find>\", \"replace\": \"\u003Cnew text>\" }\n  ]\n}\n```\n\nThe `search` text MUST be **verbatim** — character-exact. Do not summarize,\nparaphrase, or use placeholders. If the exact text is unavailable, go back to\nPhase U and gather more context before emitting.\n\nMultiple `blocks` are allowed when replacing non-adjacent hunks in the same file.\nEach block is applied in order; they MUST NOT overlap.\n\n**Form 2 — whole-file (use when the file is small or a full rewrite is cleaner):**\n\n```json\n{\n  \"target_path\": \"config.yaml\",\n  \"edit_kind\": \"whole_file\",\n  \"content\": \"\u003Cfull new file content>\"\n}\n```\n\nUse whole-file only when search\u002Freplace would require > 3 blocks or the file is\n\u003C ~80 lines. For large files, prefer targeted search\u002Freplace blocks.\n\n**Why never a diff:** small models cannot reliably apply unified diffs (Qwen-7B\n0.59 EM; disabling fuzzy apply = 9× errors — Aider\u002FDiff-XYZ studies). The harness\napplier provides deterministic fuzzy matching; Kupo's job is to emit clean\nstructured proposals.\n\n### Harness applier — TRUST IT, it is real\n\nThe nexus ships the applier. Invoke it on the scratch working copy the parent\nprovides (a throwaway copy of the repo — never the real tree):\n\n```sh\neidolons sandbox apply --proposal \u003Cproposal-json> --root \u003Cscratch-dir>\n```\n\nIt applies your search\u002Freplace (with deterministic fuzzy matching) or whole-file\nedit to the files already under `\u003Cscratch-dir>` and reports `applied` or a\nstructured error. Kupo never writes files directly — this is the only mutation path.\n\n**Do not doubt the tool.** `eidolons sandbox apply` exists and is installed. If you\ncannot immediately confirm it, RUN IT — do not assume it is missing, and do not\n`REFUSE`\u002F`ESCALATE` a KEEP task on a tool-existence doubt. A `not_found` result\nmeans your `search` text was not verbatim → return to Phase U, tighten the anchor,\nand retry. An apply error is a Phase O failure (increment counters, adjust the\nproposal, loop) — never a reason to abandon a trivially-in-scope task.\n\n### Per-file loop detector\n\nTrack attempt count per `target_path`. If the **same target file** has been\npatched **3+ times** without a green signal from Phase O:\n\n- This file is stuck. Do not keep patching.\n- Count it as 3 consecutive failures toward the circuit-breaker.\n- If the circuit-breaker has not yet fired, attempt a different approach\n  (e.g., switch edit form, re-gather context) — but only one retry.\n- If still no green signal, `ESCALATE`.\n\n---\n\n## Phase O — Observe\n\n### Running external verifiers\n\nUse the named verifier from Phase K (never substitute a different one):\n\n```sh\n# Run the NAMED verifier in the applied scratch copy (its cwd). It has no --root:\n( cd \u003Cscratch-dir> && \u003Cverifier-command> ); echo \"exit=$?\"\n# For untrusted\u002Fhost-isolated execution, wrap it via the sandbox (carries --via):\neidolons sandbox run --via \u003Csandbox-cmd> -- \u003Cverifier-command>\n```\n\nSupported verifier classes: `test`, `typecheck`, `lint`, `compile`, `diff`,\n`schema-validate`.\n\n### Success silent, failures verbose\n\n- **Green (exit 0):** proceed immediately to pre-completion gate. Do not\n  re-run; do not seek a second confirmation. One green signal suffices.\n- **Red (non-zero exit):** capture the **full** output. Keep it in context.\n  Do not truncate. Surface failures verbosely to inform the next P iteration.\n\n### Circuit-breaker (mandatory)\n\n| Counter | Threshold | Action |\n|---|---|---|\n| Consecutive failures (same or different targets) | 3 | STOP → `ESCALATE` |\n| Total failed attempts (entire session) | 20 | STOP → `ESCALATE` |\n\nA \"failure\" is any of: applier error, non-zero verifier exit, timeout,\nper-file loop detector trigger.\n\nCounters never reset within a delegation. If the circuit-breaker fires,\n`ESCALATE` immediately — do not attempt a final \"just one more\" patch.\n\n### Step ceiling and timeout\n\nRespect the host-declared step ceiling. Treat any command that exceeds the\nper-command timeout as a failure (increment both consecutive and total counters).\nDo not ignore timeouts.\n\n### Pre-completion green-signal gate\n\nEmit a `PROPOSE` **only after** ≥ 1 green external signal.\n\nThis gate is mandatory. It defeats the \"models almost never abstain\" failure mode\n(RiskEval 2601.07767) — structural, not verbal. Do not emit PROPOSE after\nintrospection (\"I'm confident this is right\") — only after the verifier exits 0.\n\n---\n\n## Output: edit-proposal artefact + ECL PROPOSE\n\nOn green signal:\n\n1. Construct the `edit-proposal` JSON matching `schemas\u002Fkupo-edit-proposal.v1.json`:\n\n   ```json\n   {\n     \"schema_version\": \"1\",\n     \"kind\": \"edit-proposal\",\n     \"task_ref\": { \"thread_id\": \"\u003Cfrom envelope>\", \"from_eidolon\": \"\u003Csender>\" },\n     \"edits\": [ { \"target_path\": \"...\", \"edit_kind\": \"search_replace\", \"blocks\": [...] } ],\n     \"verification\": {\n       \"verifier\": \"\u003Ccommand>\",\n       \"verifier_class\": \"\u003Cclass>\",\n       \"result\": \"green\",\n       \"output_excerpt\": \"\u003Cfirst 200 chars of output, or empty>\"\n     },\n     \"sandbox\": {\n       \"applied\": true,\n       \"ephemeral\": true,\n       \"applier\": \"eidolons sandbox apply\",\n       \"attempts\": \u003CN>\n     }\n   }\n   ```\n\n2. Compute `sha256` + `size_bytes` of the artefact file.\n   (If Bash is unavailable, emit `PARENT_FILLS_SHA256` and\n   `PARENT_FILLS_SIZE_BYTES` as placeholders; the parent patches these before\n   running `harness_verify`.)\n\n3. Compose the ECL envelope sidecar, validated against `schemas\u002Fecl-envelope.v2.json`\n   (`schemas\u002Fecl-envelope.v1.json` is retained in-repo for the ECL §7.3\n   back-compat window — do not emit against it for new proposals):\n\n   ```json\n   {\n     \"envelope_version\": \"2.0\",\n     \"message_id\": \"\u003CUUIDv7>\",\n     \"thread_id\": \"\u003Cfrom inbound envelope>\",\n     \"from\": { \"eidolon\": \"kupo\", \"version\": \"\u003Cversion>\" },\n     \"to\": { \"eidolon\": \"\u003Coriginal sender>\" },\n     \"performative\": \"PROPOSE\",\n     \"objective\": \"\u003C≤240-char summary of the edit>\",\n     \"artifact\": {\n       \"kind\": \"edit-proposal\",\n       \"schema_version\": \"1\",\n       \"path\": \"\u003Crelative path to artefact>\",\n       \"sha256\": \"\u003Chex>\",\n       \"size_bytes\": \u003CN>\n     },\n     \"ise\": {\n       \"assertion_grade\": \"validated\",\n       \"provenance\": {\n         \"methodology_version\": \"kupo-\u003Cversion>\",\n         \"tool_surface\": [\"Read\", \"Grep\", \"Glob\", \"Bash(eidolons sandbox:*)\"],\n         \"lateral_consults\": []\n       },\n       \"receiver_authorization\": {\n         \"auto_route\": true,\n         \"auto_merge\": false,\n         \"auto_deploy\": false\n       }\n     },\n     \"integrity\": { \"method\": \"sha256\", \"value\": \"\u003Chex>\" }\n   }\n   ```\n\n   **Why `ise.assertion_grade: \"validated\"` (ECL v2.0 §6.5) on every PROPOSE,\n   with no lower grade ever emitted:** Kupo never reaches PROPOSE except\n   through the pre-completion green-signal gate above — a NAMED external\n   verifier (the one `skills\u002Fkeep-or-kick\u002FSKILL.md` required to exist before KEEP\n   was even granted) exited green in the sandbox. That is \"emitter ran\n   spec-mandated gates\" by construction, not a self-report — `validated` is\n   earned, not defaulted. **`ise.receiver_authorization.auto_merge: false` is\n   load-bearing, not boilerplate:** the parent applies and commits (§4,\n   PROPOSE-only P0) — Kupo's own verified signal never authorizes a receiver\n   to merge on Kupo's say-so. `auto_route: true` \u002F `auto_deploy: false`\n   follow the same §6.5 defaults every Eidolon in the roster uses.\n\n4. Write both files. Signal the parent that PROPOSE is ready.\n\n5. **Post-flight procedural-memory commit (SHOULD, additive).** After the\n   green signal that gates step 1, and before or alongside emitting PROPOSE,\n   commit the verified edit pattern — see `SPEC.md §9` \"Post-flight commit\n   (KEEP verified)\". Graceful-skip when CRYSTALIUM is absent.\n\n---\n\n## Notes\n\n- **One green signal is enough.** Do not seek multiple confirmations.\n- **Keep failures in context.** Prior error output informs the next patch.\n  Do not discard failure output between P→O iterations.\n- **Escalate loudly.** When the circuit-breaker fires, include the last failure\n  output in the ESCALATE body so the parent can diagnose without re-running.\n- **Never write the real tree.** The only mutation path is\n  `eidolons sandbox apply`. Any direct file write would violate §4.\n\n---\n\n*Patch-Verify Skill — Kupo Phase P+O loop*\n","# Patch-Verify Skill — Kupo (Phase P+O loop)\n\n## When to use\n\nLoad during Phase P (Patch) and Phase O (Observe). Governs the\nedit → harness-apply → sandbox → verify loop. Never invoked in K or U.\n\nLoaded during Phase P (Patch) and Phase O (Observe). Governs the\nedit → harness-apply → sandbox → verify loop. Never invoked in K or U.\n\n---\n\n## Phase P — Patch\n\n### Edit emission\n\nEmit the edit in one of two forms. Never emit a diff.\n\n**Form 1 — search\u002Freplace (default):**\n\n```json\n{\n  \"target_path\": \"src\u002Ffoo.ts\",\n  \"edit_kind\": \"search_replace\",\n  \"blocks\": [\n    { \"search\": \"\u003Cexact verbatim text to find>\", \"replace\": \"\u003Cnew text>\" }\n  ]\n}\n```\n\nThe `search` text MUST be **verbatim** — character-exact. Do not summarize,\nparaphrase, or use placeholders. If the exact text is unavailable, go back to\nPhase U and gather more context before emitting.\n\nMultiple `blocks` are allowed when replacing non-adjacent hunks in the same file.\nEach block is applied in order; they MUST NOT overlap.\n\n**Form 2 — whole-file (use when the file is small or a full rewrite is cleaner):**\n\n```json\n{\n  \"target_path\": \"config.yaml\",\n  \"edit_kind\": \"whole_file\",\n  \"content\": \"\u003Cfull new file content>\"\n}\n```\n\nUse whole-file only when search\u002Freplace would require > 3 blocks or the file is\n\u003C ~80 lines. For large files, prefer targeted search\u002Freplace blocks.\n\n**Why never a diff:** small models cannot reliably apply unified diffs (Qwen-7B\n0.59 EM; disabling fuzzy apply = 9× errors — Aider\u002FDiff-XYZ studies). The harness\napplier provides deterministic fuzzy matching; Kupo's job is to emit clean\nstructured proposals.\n\n### Harness applier — TRUST IT, it is real\n\nThe nexus ships the applier. Invoke it on the scratch working copy the parent\nprovides (a throwaway copy of the repo — never the real tree):\n\n```sh\neidolons sandbox apply --proposal \u003Cproposal-json> --root \u003Cscratch-dir>\n```\n\nIt applies your search\u002Freplace (with deterministic fuzzy matching) or whole-file\nedit to the files already under `\u003Cscratch-dir>` and reports `applied` or a\nstructured error. Kupo never writes files directly — this is the only mutation path.\n\n**Do not doubt the tool.** `eidolons sandbox apply` exists and is installed. If you\ncannot immediately confirm it, RUN IT — do not assume it is missing, and do not\n`REFUSE`\u002F`ESCALATE` a KEEP task on a tool-existence doubt. A `not_found` result\nmeans your `search` text was not verbatim → return to Phase U, tighten the anchor,\nand retry. An apply error is a Phase O failure (increment counters, adjust the\nproposal, loop) — never a reason to abandon a trivially-in-scope task.\n\n### Per-file loop detector\n\nTrack attempt count per `target_path`. If the **same target file** has been\npatched **3+ times** without a green signal from Phase O:\n\n- This file is stuck. Do not keep patching.\n- Count it as 3 consecutive failures toward the circuit-breaker.\n- If the circuit-breaker has not yet fired, attempt a different approach\n  (e.g., switch edit form, re-gather context) — but only one retry.\n- If still no green signal, `ESCALATE`.\n\n---\n\n## Phase O — Observe\n\n### Running external verifiers\n\nUse the named verifier from Phase K (never substitute a different one):\n\n```sh\n# Run the NAMED verifier in the applied scratch copy (its cwd). It has no --root:\n( cd \u003Cscratch-dir> && \u003Cverifier-command> ); echo \"exit=$?\"\n# For untrusted\u002Fhost-isolated execution, wrap it via the sandbox (carries --via):\neidolons sandbox run --via \u003Csandbox-cmd> -- \u003Cverifier-command>\n```\n\nSupported verifier classes: `test`, `typecheck`, `lint`, `compile`, `diff`,\n`schema-validate`.\n\n### Success silent, failures verbose\n\n- **Green (exit 0):** proceed immediately to pre-completion gate. Do not\n  re-run; do not seek a second confirmation. One green signal suffices.\n- **Red (non-zero exit):** capture the **full** output. Keep it in context.\n  Do not truncate. Surface failures verbosely to inform the next P iteration.\n\n### Circuit-breaker (mandatory)\n\n| Counter | Threshold | Action |\n|---|---|---|\n| Consecutive failures (same or different targets) | 3 | STOP → `ESCALATE` |\n| Total failed attempts (entire session) | 20 | STOP → `ESCALATE` |\n\nA \"failure\" is any of: applier error, non-zero verifier exit, timeout,\nper-file loop detector trigger.\n\nCounters never reset within a delegation. If the circuit-breaker fires,\n`ESCALATE` immediately — do not attempt a final \"just one more\" patch.\n\n### Step ceiling and timeout\n\nRespect the host-declared step ceiling. Treat any command that exceeds the\nper-command timeout as a failure (increment both consecutive and total counters).\nDo not ignore timeouts.\n\n### Pre-completion green-signal gate\n\nEmit a `PROPOSE` **only after** ≥ 1 green external signal.\n\nThis gate is mandatory. It defeats the \"models almost never abstain\" failure mode\n(RiskEval 2601.07767) — structural, not verbal. Do not emit PROPOSE after\nintrospection (\"I'm confident this is right\") — only after the verifier exits 0.\n\n---\n\n## Output: edit-proposal artefact + ECL PROPOSE\n\nOn green signal:\n\n1. Construct the `edit-proposal` JSON matching `schemas\u002Fkupo-edit-proposal.v1.json`:\n\n   ```json\n   {\n     \"schema_version\": \"1\",\n     \"kind\": \"edit-proposal\",\n     \"task_ref\": { \"thread_id\": \"\u003Cfrom envelope>\", \"from_eidolon\": \"\u003Csender>\" },\n     \"edits\": [ { \"target_path\": \"...\", \"edit_kind\": \"search_replace\", \"blocks\": [...] } ],\n     \"verification\": {\n       \"verifier\": \"\u003Ccommand>\",\n       \"verifier_class\": \"\u003Cclass>\",\n       \"result\": \"green\",\n       \"output_excerpt\": \"\u003Cfirst 200 chars of output, or empty>\"\n     },\n     \"sandbox\": {\n       \"applied\": true,\n       \"ephemeral\": true,\n       \"applier\": \"eidolons sandbox apply\",\n       \"attempts\": \u003CN>\n     }\n   }\n   ```\n\n2. Compute `sha256` + `size_bytes` of the artefact file.\n   (If Bash is unavailable, emit `PARENT_FILLS_SHA256` and\n   `PARENT_FILLS_SIZE_BYTES` as placeholders; the parent patches these before\n   running `harness_verify`.)\n\n3. Compose the ECL envelope sidecar, validated against `schemas\u002Fecl-envelope.v2.json`\n   (`schemas\u002Fecl-envelope.v1.json` is retained in-repo for the ECL §7.3\n   back-compat window — do not emit against it for new proposals):\n\n   ```json\n   {\n     \"envelope_version\": \"2.0\",\n     \"message_id\": \"\u003CUUIDv7>\",\n     \"thread_id\": \"\u003Cfrom inbound envelope>\",\n     \"from\": { \"eidolon\": \"kupo\", \"version\": \"\u003Cversion>\" },\n     \"to\": { \"eidolon\": \"\u003Coriginal sender>\" },\n     \"performative\": \"PROPOSE\",\n     \"objective\": \"\u003C≤240-char summary of the edit>\",\n     \"artifact\": {\n       \"kind\": \"edit-proposal\",\n       \"schema_version\": \"1\",\n       \"path\": \"\u003Crelative path to artefact>\",\n       \"sha256\": \"\u003Chex>\",\n       \"size_bytes\": \u003CN>\n     },\n     \"ise\": {\n       \"assertion_grade\": \"validated\",\n       \"provenance\": {\n         \"methodology_version\": \"kupo-\u003Cversion>\",\n         \"tool_surface\": [\"Read\", \"Grep\", \"Glob\", \"Bash(eidolons sandbox:*)\"],\n         \"lateral_consults\": []\n       },\n       \"receiver_authorization\": {\n         \"auto_route\": true,\n         \"auto_merge\": false,\n         \"auto_deploy\": false\n       }\n     },\n     \"integrity\": { \"method\": \"sha256\", \"value\": \"\u003Chex>\" }\n   }\n   ```\n\n   **Why `ise.assertion_grade: \"validated\"` (ECL v2.0 §6.5) on every PROPOSE,\n   with no lower grade ever emitted:** Kupo never reaches PROPOSE except\n   through the pre-completion green-signal gate above — a NAMED external\n   verifier (the one `skills\u002Fkeep-or-kick\u002FSKILL.md` required to exist before KEEP\n   was even granted) exited green in the sandbox. That is \"emitter ran\n   spec-mandated gates\" by construction, not a self-report — `validated` is\n   earned, not defaulted. **`ise.receiver_authorization.auto_merge: false` is\n   load-bearing, not boilerplate:** the parent applies and commits (§4,\n   PROPOSE-only P0) — Kupo's own verified signal never authorizes a receiver\n   to merge on Kupo's say-so. `auto_route: true` \u002F `auto_deploy: false`\n   follow the same §6.5 defaults every Eidolon in the roster uses.\n\n4. Write both files. Signal the parent that PROPOSE is ready.\n\n5. **Post-flight procedural-memory commit (SHOULD, additive).** After the\n   green signal that gates step 1, and before or alongside emitting PROPOSE,\n   commit the verified edit pattern — see `SPEC.md §9` \"Post-flight commit\n   (KEEP verified)\". Graceful-skip when CRYSTALIUM is absent.\n\n---\n\n## Notes\n\n- **One green signal is enough.** Do not seek multiple confirmations.\n- **Keep failures in context.** Prior error output informs the next patch.\n  Do not discard failure output between P→O iterations.\n- **Escalate loudly.** When the circuit-breaker fires, include the last failure\n  output in the ESCALATE body so the parent can diagnose without re-running.\n- **Never write the real tree.** The only mutation path is\n  `eidolons sandbox apply`. Any direct file write would violate §4.\n\n---\n\n*Patch-Verify Skill — Kupo Phase P+O loop*\n",{"name":5,"metadata":24,"description":7},{"phase":25,"methodology":26},"P+O","Kupo","main","38c3a82583abf9806de764f89ac0099ad690933a",".eidolons\u002Fkupo\u002Fskills\u002Fpatch-verify",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":37,"prompt":47,"zipAvailable":48},[38,41,44],{"tool":39,"command":40},"npx","npx skills add Rynaro\u002Fmagicite --skill kupo-patch-verify",{"tool":42,"command":43},"bunx","bunx skills add Rynaro\u002Fmagicite --skill kupo-patch-verify",{"tool":45,"command":46},"pnpm","pnpm dlx skills add Rynaro\u002Fmagicite --skill kupo-patch-verify","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FRynaro\u002Fmagicite\u002Fblob\u002Fmain\u002F.eidolons\u002Fkupo\u002Fskills\u002Fpatch-verify\u002FSKILL.md",true,[],[51,58,65,72,79,86],{"skillNo":52,"name":53,"nameZh":6,"description":54,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":55,"categories":56,"keywords":57,"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":59,"name":60,"nameZh":6,"description":61,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":62,"categories":63,"keywords":64,"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":66,"name":67,"nameZh":6,"description":68,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":69,"categories":70,"keywords":71,"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":73,"name":74,"nameZh":6,"description":75,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":76,"categories":77,"keywords":78,"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":80,"name":81,"nameZh":6,"description":82,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":83,"categories":84,"keywords":85,"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":87,"name":88,"nameZh":6,"description":89,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":90,"categories":91,"keywords":92,"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",[],[]]