[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M78GPYNQ9D9PKGYJPRS6TC73":3,"skill-same:01M78GPYNQ9D9PKGYJPRS6TC73":47},{"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":30,"ownerDescription":31,"updatedAt":32,"install":33,"resources":46},"01M78GPYNQ9D9PKGYJPRS6TC73","pr-patrol",null,"Use when the user wants to review someone else's PR, says 'review this PR', 'look at this PR', 'check this PR', shares a PR URL with review intent, or asks to post PR review comments. Fetches the PR diff, runs the review contract, drafts safe comments, and posts only after approval.","sandsower\u002Fnopal","sandsower","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F302322?v=4","01M78GNBK4G12AGP2PA2FND7R4","Vic Valenzuela","01M78GNBM80RBPX57WYVHFHC97",0,"MIT","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fpr-patrol\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: pr-patrol\ndescription: \"Use when the user wants to review someone else's PR, says 'review this PR', 'look at this PR', 'check this PR', shares a PR URL with review intent, or asks to post PR review comments. Fetches the PR diff, runs the review contract, drafts safe comments, and posts only after approval.\"\n---\n\n# PR Patrol\n\nReview someone else's pull request and optionally post approved review comments. This is an inbound review workflow built around the side-effect-free `review` contract.\n\nBoundary:\n- `review` inspects a diff and returns findings.\n- `pr-patrol` fetches PR context, decides with the user which findings are worth posting, drafts comments, and performs posting side effects only after approval.\n\nUse this for:\n- Reviewing another person's GitHub\u002FGitLab\u002Fetc. PR\n- Turning review-contract findings into human-approved PR comments\n- Running a fresh-eyes pass on a remote PR before posting comments\n- Security-sensitive inbound reviews where multiple reviewers may be worth the cost\n\nDo not use this for:\n- Reviewing your own local diff before PR handoff — use `review`, `fresh-eyes`, or `ready-for-review`\n- Addressing feedback on your own PR — use `review-response`\n- Interactive walkthroughs of your own changes — use `walk-the-diff`\n- Posting comments without explicit user approval\n\n## Preflight\n\nBefore touching external systems, establish capabilities and print them to the user.\n\nResolve:\n- PR URL or host\u002Fowner\u002Frepo\u002Fnumber. **If the user did not supply one, default to the PR for the current branch** — try `gh pr view --json number,url,author,headRefName,state` (or the host equivalent) before asking. Only prompt the user if no PR is associated with the current branch, or if the current branch's author is the user themselves (pr-patrol is for inbound review, not your own PRs).\n- PR host tooling available in this session (`gh`, `glab`, MCP tool, API client, or paste-only)\n- Whether inline comments are supported by the available tool\n- Whether the user wants comment drafts only or actual posting\n\nExample capability table:\n\n```text\npr-patrol preflight\n  pr:              \u003Curl>\n  host:            github \u002F gitlab \u002F other \u002F paste\n  fetch_diff:      gh \u002F glab \u002F mcp \u002F paste        ✓\n  inline_comments: yes\u002Fno                         ✓\n  post_comments:   yes\u002Fno                         ✓\n```\n\nIf required tooling is missing, offer:\n\n```text\n(a) load\u002Ffix tooling and retry\n(b) continue in draft-only mode\n(c) abort\n```\n\nDo not post anything unless posting is explicitly available and the user approves the final drafts.\n\n## Checklist\n\nComplete these phases in order:\n\n1. Fetch PR metadata and diff\n2. Load requirements\u002Fcontext\n3. Run the review contract\n4. Select findings to post\n5. Draft comments\n6. Post approved comments, or print drafts\n7. Summarize review outcome\n\n## Phase 1: Fetch PR metadata and diff\n\nAccept a PR URL or host-specific identifier. Prefer host tooling when available.\n\nFor GitHub with `gh` available:\n\n```bash\ngh pr view \u003Curl-or-number> --json title,body,author,baseRefName,headRefName,url,files\nmkdir -p \"${TMPDIR:-\u002Ftmp}\u002Fbeislid-pr-patrol\"\ngh pr diff \u003Curl-or-number> > \"${TMPDIR:-\u002Ftmp}\u002Fbeislid-pr-patrol\u002Fpr.diff\"\n```\n\nFor other hosts, use the configured MCP\u002Ftooling or ask the user to paste:\n- PR title\n- PR description\n- Base\u002Fhead branches\n- Diff or patch\n- Any relevant ticket\u002Fspec context\n\nIf the PR is too large for one review pass, split by logical area and say how findings will be grouped. Do not silently ignore files.\n\n## Phase 2: Load requirements\u002Fcontext\n\nUse, in order:\n- PR title and description\n- Linked issue\u002Fticket\u002Fspec in the PR body\n- Caller-provided focus areas\n- Commit messages if available\n- General production readiness if requirements are missing\n\nIf requirements are unavailable, say so in the review metadata.\n\n## Phase 3: Run review\n\nInvoke the `review` contract with:\n- PR metadata\n- Requirements\u002Fcontext\n- Full diff or scoped diff chunk\n- Focus areas from the user\n\nFor large PRs, run one overall pass plus scoped passes by subsystem when possible.\n\nFor security-sensitive PRs — auth, permissions, token handling, crypto, secrets, privacy, or data export — offer multi-reviewer mode if the host can call additional independent reviewers. All reviewers must return the same review contract and must not post comments.\n\nOptional: run `fresh-eyes` after the normal review for whole-diff drift, especially on large PRs.\n\n## Phase 4: Select findings to post\n\nShow findings grouped by severity. Recommend posting:\n- Critical findings\n- Important findings with concrete evidence and a clear requested change\n- Minor findings only when they are actionable and worth reviewer attention\n\nAsk the user which findings to post:\n\n```text\nWhich findings should I post?\n(a) all Critical + Important\n(b) only selected IDs: \u003CIDs>\n(c) draft comments only; do not post\n(d) none; summarize locally\n```\n\nDo not post style-only nits unless the user explicitly selects them.\n\n## Phase 5: Draft comments\n\nConvert selected findings into concise PR comments. Each comment should include:\n- The issue\n- The evidence\n- Why it matters\n- A concrete suggested fix\n\nTone:\n- Direct, specific, and kind\n- No performative hedging\n- No accusations\n- No generic AI phrasing\n\nWhen inline comments are supported, map each finding to a file\u002Fpath\u002Fline. If line mapping is unavailable, use a general PR review comment with file references.\n\n### Safe comment body handling\n\nUse temporary files for comment bodies. Do not inline large comment text in shell commands. This avoids shell quoting bugs and credential-guard false positives.\n\nExample pattern:\n\n```bash\ncomment_file=\"$(mktemp \"${TMPDIR:-\u002Ftmp}\u002Fbeislid-pr-comment.XXXXXX.md\")\"\n# Write the approved body to \"$comment_file\" with the host agent's file tool;\n# never open an interactive editor from a non-interactive agent session.\n# Pass the file as a body field if the host tool supports file-backed fields.\n# Otherwise build a JSON payload file and pass that file to the API client.\nrm -f \"$comment_file\"\n```\n\nIf the host\u002FAPI requires a JSON payload, write JSON to a temp file and pass the payload file path to the tool instead of embedding the body in the command string. Do not pass a raw Markdown body file to an API option that expects JSON.\n\n## Phase 6: Approval gate and posting\n\nPresent all drafts before posting as context, then ask the explicit approval question once in the final response:\n\n```md\n### Comment Drafts\n#### P1 -> \u003Cpath:line or general>\n\u003Cbody>\n...\n```\n\nAsk once in the final response:\n\n```text\nPost these comments? yes\u002Fno\u002Fedit\n```\n\n- On **yes**: post only the approved comments.\n- On **edit**: update drafts, then ask again.\n- On **no**: do not post; print the drafts for manual use if requested.\n\nFor GitHub, prefer host-supported review APIs. If the exact inline position cannot be resolved safely, fall back to a general PR review comment rather than guessing an inline location.\n\n## Phase 7: Summarize\n\nEnd with:\n\n```md\n### PR Patrol Summary\n- PR: \u003Curl>\n- Requirements: \u003Csource or not available>\n- Reviewers used: local \u002F external names if any\n- Findings found: Critical N, Important N, Minor N\n- Comments posted: N\n- Draft-only comments: N\n- Not posted: \u003CIDs and reason>\n- Suggested review verdict: approve \u002F comment \u002F request changes\n```\n\n## Common mistakes\n\n- **Posting without approval** — every comment needs explicit user approval.\n- **Letting review post comments** — `review` is side-effect-free. `pr-patrol` owns posting.\n- **Shell-inlining comment bodies** — use temp files for comment bodies and JSON payloads.\n- **Guessing inline positions** — fall back to general comments when mapping is uncertain.\n- **Overposting minor nits** — inbound review should be high signal.\n- **Hardcoding one tenant or repo host** — prefer host-agnostic wording; use `gh` examples only as examples.\n","# PR Patrol\n\nReview someone else's pull request and optionally post approved review comments. This is an inbound review workflow built around the side-effect-free `review` contract.\n\nBoundary:\n- `review` inspects a diff and returns findings.\n- `pr-patrol` fetches PR context, decides with the user which findings are worth posting, drafts comments, and performs posting side effects only after approval.\n\nUse this for:\n- Reviewing another person's GitHub\u002FGitLab\u002Fetc. PR\n- Turning review-contract findings into human-approved PR comments\n- Running a fresh-eyes pass on a remote PR before posting comments\n- Security-sensitive inbound reviews where multiple reviewers may be worth the cost\n\nDo not use this for:\n- Reviewing your own local diff before PR handoff — use `review`, `fresh-eyes`, or `ready-for-review`\n- Addressing feedback on your own PR — use `review-response`\n- Interactive walkthroughs of your own changes — use `walk-the-diff`\n- Posting comments without explicit user approval\n\n## Preflight\n\nBefore touching external systems, establish capabilities and print them to the user.\n\nResolve:\n- PR URL or host\u002Fowner\u002Frepo\u002Fnumber. **If the user did not supply one, default to the PR for the current branch** — try `gh pr view --json number,url,author,headRefName,state` (or the host equivalent) before asking. Only prompt the user if no PR is associated with the current branch, or if the current branch's author is the user themselves (pr-patrol is for inbound review, not your own PRs).\n- PR host tooling available in this session (`gh`, `glab`, MCP tool, API client, or paste-only)\n- Whether inline comments are supported by the available tool\n- Whether the user wants comment drafts only or actual posting\n\nExample capability table:\n\n```text\npr-patrol preflight\n  pr:              \u003Curl>\n  host:            github \u002F gitlab \u002F other \u002F paste\n  fetch_diff:      gh \u002F glab \u002F mcp \u002F paste        ✓\n  inline_comments: yes\u002Fno                         ✓\n  post_comments:   yes\u002Fno                         ✓\n```\n\nIf required tooling is missing, offer:\n\n```text\n(a) load\u002Ffix tooling and retry\n(b) continue in draft-only mode\n(c) abort\n```\n\nDo not post anything unless posting is explicitly available and the user approves the final drafts.\n\n## Checklist\n\nComplete these phases in order:\n\n1. Fetch PR metadata and diff\n2. Load requirements\u002Fcontext\n3. Run the review contract\n4. Select findings to post\n5. Draft comments\n6. Post approved comments, or print drafts\n7. Summarize review outcome\n\n## Phase 1: Fetch PR metadata and diff\n\nAccept a PR URL or host-specific identifier. Prefer host tooling when available.\n\nFor GitHub with `gh` available:\n\n```bash\ngh pr view \u003Curl-or-number> --json title,body,author,baseRefName,headRefName,url,files\nmkdir -p \"${TMPDIR:-\u002Ftmp}\u002Fbeislid-pr-patrol\"\ngh pr diff \u003Curl-or-number> > \"${TMPDIR:-\u002Ftmp}\u002Fbeislid-pr-patrol\u002Fpr.diff\"\n```\n\nFor other hosts, use the configured MCP\u002Ftooling or ask the user to paste:\n- PR title\n- PR description\n- Base\u002Fhead branches\n- Diff or patch\n- Any relevant ticket\u002Fspec context\n\nIf the PR is too large for one review pass, split by logical area and say how findings will be grouped. Do not silently ignore files.\n\n## Phase 2: Load requirements\u002Fcontext\n\nUse, in order:\n- PR title and description\n- Linked issue\u002Fticket\u002Fspec in the PR body\n- Caller-provided focus areas\n- Commit messages if available\n- General production readiness if requirements are missing\n\nIf requirements are unavailable, say so in the review metadata.\n\n## Phase 3: Run review\n\nInvoke the `review` contract with:\n- PR metadata\n- Requirements\u002Fcontext\n- Full diff or scoped diff chunk\n- Focus areas from the user\n\nFor large PRs, run one overall pass plus scoped passes by subsystem when possible.\n\nFor security-sensitive PRs — auth, permissions, token handling, crypto, secrets, privacy, or data export — offer multi-reviewer mode if the host can call additional independent reviewers. All reviewers must return the same review contract and must not post comments.\n\nOptional: run `fresh-eyes` after the normal review for whole-diff drift, especially on large PRs.\n\n## Phase 4: Select findings to post\n\nShow findings grouped by severity. Recommend posting:\n- Critical findings\n- Important findings with concrete evidence and a clear requested change\n- Minor findings only when they are actionable and worth reviewer attention\n\nAsk the user which findings to post:\n\n```text\nWhich findings should I post?\n(a) all Critical + Important\n(b) only selected IDs: \u003CIDs>\n(c) draft comments only; do not post\n(d) none; summarize locally\n```\n\nDo not post style-only nits unless the user explicitly selects them.\n\n## Phase 5: Draft comments\n\nConvert selected findings into concise PR comments. Each comment should include:\n- The issue\n- The evidence\n- Why it matters\n- A concrete suggested fix\n\nTone:\n- Direct, specific, and kind\n- No performative hedging\n- No accusations\n- No generic AI phrasing\n\nWhen inline comments are supported, map each finding to a file\u002Fpath\u002Fline. If line mapping is unavailable, use a general PR review comment with file references.\n\n### Safe comment body handling\n\nUse temporary files for comment bodies. Do not inline large comment text in shell commands. This avoids shell quoting bugs and credential-guard false positives.\n\nExample pattern:\n\n```bash\ncomment_file=\"$(mktemp \"${TMPDIR:-\u002Ftmp}\u002Fbeislid-pr-comment.XXXXXX.md\")\"\n# Write the approved body to \"$comment_file\" with the host agent's file tool;\n# never open an interactive editor from a non-interactive agent session.\n# Pass the file as a body field if the host tool supports file-backed fields.\n# Otherwise build a JSON payload file and pass that file to the API client.\nrm -f \"$comment_file\"\n```\n\nIf the host\u002FAPI requires a JSON payload, write JSON to a temp file and pass the payload file path to the tool instead of embedding the body in the command string. Do not pass a raw Markdown body file to an API option that expects JSON.\n\n## Phase 6: Approval gate and posting\n\nPresent all drafts before posting as context, then ask the explicit approval question once in the final response:\n\n```md\n### Comment Drafts\n#### P1 -> \u003Cpath:line or general>\n\u003Cbody>\n...\n```\n\nAsk once in the final response:\n\n```text\nPost these comments? yes\u002Fno\u002Fedit\n```\n\n- On **yes**: post only the approved comments.\n- On **edit**: update drafts, then ask again.\n- On **no**: do not post; print the drafts for manual use if requested.\n\nFor GitHub, prefer host-supported review APIs. If the exact inline position cannot be resolved safely, fall back to a general PR review comment rather than guessing an inline location.\n\n## Phase 7: Summarize\n\nEnd with:\n\n```md\n### PR Patrol Summary\n- PR: \u003Curl>\n- Requirements: \u003Csource or not available>\n- Reviewers used: local \u002F external names if any\n- Findings found: Critical N, Important N, Minor N\n- Comments posted: N\n- Draft-only comments: N\n- Not posted: \u003CIDs and reason>\n- Suggested review verdict: approve \u002F comment \u002F request changes\n```\n\n## Common mistakes\n\n- **Posting without approval** — every comment needs explicit user approval.\n- **Letting review post comments** — `review` is side-effect-free. `pr-patrol` owns posting.\n- **Shell-inlining comment bodies** — use temp files for comment bodies and JSON payloads.\n- **Guessing inline positions** — fall back to general comments when mapping is uncertain.\n- **Overposting minor nits** — inbound review should be high signal.\n- **Hardcoding one tenant or repo host** — prefer host-agnostic wording; use `gh` examples only as examples.\n",{"name":5,"description":7},"main","bba9b487e6731261bcfe35e17ea03269cf8df0a1","resources\u002Fbeislid\u002Fskills\u002Fpr-patrol","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal","https:\u002F\u002Fnopal.sh","https:\u002F\u002Fgithub.com\u002Fsandsower","vic.valenzuela.is","Product-minded software engineer","2026-10-15 00:33:24",{"source":8,"skillName":5,"commands":34,"prompt":44,"zipAvailable":45},[35,38,41],{"tool":36,"command":37},"npx","npx skills add sandsower\u002Fnopal --skill pr-patrol",{"tool":39,"command":40},"bunx","bunx skills add sandsower\u002Fnopal --skill pr-patrol",{"tool":42,"command":43},"pnpm","pnpm dlx skills add sandsower\u002Fnopal --skill pr-patrol","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fpr-patrol\u002FSKILL.md",true,[],[48,55,62,69,76,83],{"skillNo":49,"name":50,"nameZh":6,"description":51,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":52,"categories":53,"keywords":54,"publishedAt":19},"01M78GRB8RYKHHGYZF3FGN9Y5P","walk-the-diff","Use when the user wants to walk through their local changes interactively before or instead of a PR review — e.g., \"guided review\", \"walk me through the changes\", \"review my branch\", \"let's go through the diff\". Also use when the user explicitly invokes the walk-the-diff skill through the host agent. This is for interactive human-facing walkthroughs of code changes, not automated code review.","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fwalk-the-diff\u002FSKILL.md",[],[],{"skillNo":56,"name":57,"nameZh":6,"description":58,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":59,"categories":60,"keywords":61,"publishedAt":19},"01M78GR8YWH8JMB2E9RRSAR7SF","verify","Use before claiming work is complete, fixed, or passing. Requires running verification commands and reading output before making any success claims. Evidence before assertions.","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fverify\u002FSKILL.md",[],[],{"skillNo":63,"name":64,"nameZh":6,"description":65,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":19},"01M78GR26XMDX0KAF5MYZ2W0TN","spec","Trigger immediately when the user says any of: 'spec it out', 'spec this out', 'before you build', 'before we build', 'before implementation', 'let's discuss before building', 'discuss and spec', 'PRD this', 'shape this', 'product spec', 'lightweight spec'. Also use for brainstorming, vague tickets, product shaping, or writing a lightweight product spec\u002FPRD when the problem, scope, desired behavior, or success criteria are unclear. Do not skip this skill and answer conversationally yourself — if a trigger phrase appears, route here even if the request seems clear or auto mode is active. Do not use for implementation design when requirements are already clear; use blueprint instead.","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fspec\u002FSKILL.md",[],[],{"skillNo":70,"name":71,"nameZh":6,"description":72,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":73,"categories":74,"keywords":75,"publishedAt":19},"01M78GQXP7EJ8WHHZJ4H1NNN0N","show-me","Use when the user asks for a visual HTML evidence\u002Fexplanation artifact: 'show-me', 'show me this', 'make a proof deck', 'make an HTML deck', 'visualize this as HTML', 'this would be better as slides', 'make this visual', 'create a walkthrough deck', 'document this visually', or 'review this visually'. Manual\u002Fuser-requested only in v1. Produces a polished local HTML portfolio\u002Fdeck with evidence, media, logs, provenance, and clear missing-evidence markers.","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fshow-me\u002FSKILL.md",[],[],{"skillNo":77,"name":78,"nameZh":6,"description":79,"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},"01M78GQS6VPCC70BNE86HSAF6S","setup","Configure Beislið project config (`.beislid\u002Fworkflow.md`) interactively, or update an installed Beislið distribution. Use when the user says \"setup\", \"\u002Fsetup\", \"setup update\", \"\u002Fsetup update\", \"update beislid\", \"configure beislid\", \"set up workflow\", \"add scopes\", \"change ticket source\", \"reconfigure beislid\", or any intent to add, change, or remove a workflow.md section. Walks through sections conversationally; shows a diff before any destructive write.","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Fsetup\u002FSKILL.md",[],[],{"skillNo":84,"name":85,"nameZh":6,"description":86,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":87,"categories":88,"keywords":89,"publishedAt":19},"01M78GQNPYS736XNYDGKW5N7F4","rinse","Use when the user asks to 'rinse', 'run a review\u002Ffix loop', 'iterate on review findings', 'review and fix until clean', or harden a branch before PR handoff. Orchestrates side-effect-free review, user-approved fixes, verification, reruns, and redesign stop conditions.","https:\u002F\u002Fgithub.com\u002Fsandsower\u002Fnopal\u002Fblob\u002Fmain\u002Fresources\u002Fbeislid\u002Fskills\u002Frinse\u002FSKILL.md",[],[]]