[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M74Q1FDMBVYDVSJCDEWKT6B2":3,"skill-same:01M74Q1FDMBVYDVSJCDEWKT6B2":54},{"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":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":29,"ownerHtmlUrl":30,"ownerBlogUrl":31,"ownerDescription":32,"updatedAt":33,"install":34,"resources":47},"01M74Q1FDMBVYDVSJCDEWKT6B2","verify-green",null,"Select and run the smallest risk-based verification lane that proves a mono-agent change. Use before claiming a change works, before committing or opening a PR, or when asked to verify, run the gate, make it green, or check architecture.","robertsreberski\u002Fmono-agent","robertsreberski","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F8419292?v=4","01M74PZ37AVTNF0C6ATE0G6HV3","Robert Sreberski","01M74PZ37Y74736M79A4FTQ5AW",2,"GPL-3.0","https:\u002F\u002Fgithub.com\u002Frobertsreberski\u002Fmono-agent\u002Fblob\u002Fmain\u002Fskills\u002Fverify-green\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: verify-green\ndescription: Select and run the smallest risk-based verification lane that proves a mono-agent change. Use before claiming a change works, before committing or opening a PR, or when asked to verify, run the gate, make it green, or check architecture.\n---\n\n# Verify green\n\nVerification is selected from the diff, not accumulated from every available\nworkflow. Run focused checks while iterating, then execute the chosen lane once\nbefore shipping. Do not repeat a passing broad gate at the same SHA.\n\n## 1. Classify the change\n\nChoose the highest applicable lane:\n\n| Lane | Typical diff | Required proof |\n|---|---|---|\n| Process | `AGENTS.md`, `skills\u002F`, agent metadata, internal docs | Relevant contract tests, discoverability, OSS hygiene, whitespace |\n| Package | Ordinary implementation within existing package boundaries | Dependency-closure build, focused tests, package typecheck, one broad CI gate |\n| High risk | Security\u002Fsandbox, storage or migration, lifecycle, provider routing, delivery, public package boundaries | Focused checks, one local `verify:all`, CI, one matching live smoke |\n| Release | Mechanical lockstep version\u002Fpublish work | Use `release-lockstep`; do not add deployment or consumer checks here |\n\nA mixed diff uses the highest lane. Explain the classification in the final\nevidence so reviewers can challenge it.\n\n## 1a. Cross-process contract changes\n\nA **wire contract** is any value this repo emits that a different process parses\nlater: `callback_data` protocols, event and tool names, on-disk state shapes,\nnotification payloads, CLI JSON. Renaming or retiring one is always **high\nrisk**, however small the diff, because in-repo tests cannot observe the other\nside. Green tests are the expected result of this break, not evidence against it.\n\nBefore merging such a diff:\n\n```bash\n# in-repo consumers\ngrep -rn \"\u003Cold protocol literal>\" packages\u002F*\u002Fsrc extras\u002F*\u002Fsrc --include=\"*.ts\"\n# out-of-repo consumers — the fleet parses these contracts too\ngrep -rn \"\u003Cold protocol literal>\" \u003Cinstance-dir> \u003Cother-instance-dirs> \\\n  --include=\"*.mjs\" --include=\"*.ts\" --include=\"*.md\" 2>\u002Fdev\u002Fnull \\\n  | grep -v node_modules | grep -v \"\u002Fdist\u002F\"\n```\n\nThen require the matching `live-smoke` round-trip: emit the new contract and\nconfirm the far side acts on it. A unit test asserting a **literal** (`[\"ask:0\",\n\"ask:1\"]`) proves only that the literal is unchanged — derive the expectation\nfrom the exported builder instead, so retiring the protocol fails the test that\npins it.\n\nPR #527 renamed the Telegram button protocol `ask:\u003Cindex>` → `reply:v1:\u003C0-7>`\nwith every gate green. Personal Agent heartbeat cards kept emitting the old\npayload for a week; taps did nothing, and a test pinning the literal stayed green\nthroughout.\n\n## 2. Process lane\n\nDo not build packages, run live smoke, inspect the fleet, or run release checks\nfor a docs\u002Fskills\u002Fprocess-only diff. Select tests that cover the edited contract.\nFor the engineering-skill surface, use:\n\n```bash\npnpm exec vitest run \\\n  scripts\u002F__tests__\u002Fcheck-codex-discoverability.test.mjs \\\n  scripts\u002F__tests__\u002Frepo-hygiene-skills.test.mjs \\\n  scripts\u002Frelease\u002F__tests__\u002Fpackage-count-drift.test.mjs\npnpm run check:codex-discoverability\npnpm run check:oss-hygiene\ngit diff --check\n```\n\nRun the website build only when `docs\u002F`, `website\u002F`, or published-site inputs\nchanged; use `docs-sync` to choose that surface.\n\n## 3. Package lane\n\nWhile iterating on package X:\n\n```bash\npnpm --filter @mono-agent\u002F\u003CX>... run build\npnpm --filter @mono-agent\u002F\u003CX> test\npnpm --filter @mono-agent\u002F\u003CX> run typecheck\n```\n\nAdd focused dependent-package checks when the changed public behavior crosses a\npackage boundary. Before merge, rely on one successful hosted CI run for the\nexact head SHA. If Actions is unavailable, run `pnpm run verify:all` once under\na supported Node runtime and record why CI could not provide the broad gate.\n\n## 4. High-risk lane\n\nRun the focused package checks above, then:\n\n```bash\npnpm run verify:all\n```\n\nRequire hosted CI for the exact head SHA and select exactly one scenario from\n`live-smoke` that exercises the risky boundary. Add a second scenario only when\nthe diff independently changes a second live surface.\n\nProvider-backed smoke is reserved for provider behavior. UI, worker transport,\nand lifecycle changes should use their narrower local scenarios instead of paid\nmodel calls.\n\n## Worktree dist rule\n\nCross-package imports resolve through built `dist\u002F`. After editing package A,\nrebuild A before testing or typechecking a dependent package. A process-only\nchange does not exercise package resolution and needs no dist baseline. See\n`worktree-feature` for diff-aware worktree setup.\n\n## Failure handling\n\n- Diagnose the first stable failure; do not rerun an unchanged failing broad gate.\n- When a failure may pre-exist, compare against the exact base SHA in a detached worktree.\n- A failure that passes on rerun at the same SHA is a flake: route it with `flaky-tests` (fix in scope, issue otherwise) instead of only rerunning.\n- If hosted CI fails immediately for an account or billing condition, stop polling and report that external blocker.\n- Never replace a failed stated check with a different check without naming the substitution.\n\n## Report format\n\nState the lane, exact commands, SHA, and pass\u002Ffail result. For failures, name the\nfailing test or gate. A green claim without command evidence is not complete.\n","# Verify green\n\nVerification is selected from the diff, not accumulated from every available\nworkflow. Run focused checks while iterating, then execute the chosen lane once\nbefore shipping. Do not repeat a passing broad gate at the same SHA.\n\n## 1. Classify the change\n\nChoose the highest applicable lane:\n\n| Lane | Typical diff | Required proof |\n|---|---|---|\n| Process | `AGENTS.md`, `skills\u002F`, agent metadata, internal docs | Relevant contract tests, discoverability, OSS hygiene, whitespace |\n| Package | Ordinary implementation within existing package boundaries | Dependency-closure build, focused tests, package typecheck, one broad CI gate |\n| High risk | Security\u002Fsandbox, storage or migration, lifecycle, provider routing, delivery, public package boundaries | Focused checks, one local `verify:all`, CI, one matching live smoke |\n| Release | Mechanical lockstep version\u002Fpublish work | Use `release-lockstep`; do not add deployment or consumer checks here |\n\nA mixed diff uses the highest lane. Explain the classification in the final\nevidence so reviewers can challenge it.\n\n## 1a. Cross-process contract changes\n\nA **wire contract** is any value this repo emits that a different process parses\nlater: `callback_data` protocols, event and tool names, on-disk state shapes,\nnotification payloads, CLI JSON. Renaming or retiring one is always **high\nrisk**, however small the diff, because in-repo tests cannot observe the other\nside. Green tests are the expected result of this break, not evidence against it.\n\nBefore merging such a diff:\n\n```bash\n# in-repo consumers\ngrep -rn \"\u003Cold protocol literal>\" packages\u002F*\u002Fsrc extras\u002F*\u002Fsrc --include=\"*.ts\"\n# out-of-repo consumers — the fleet parses these contracts too\ngrep -rn \"\u003Cold protocol literal>\" \u003Cinstance-dir> \u003Cother-instance-dirs> \\\n  --include=\"*.mjs\" --include=\"*.ts\" --include=\"*.md\" 2>\u002Fdev\u002Fnull \\\n  | grep -v node_modules | grep -v \"\u002Fdist\u002F\"\n```\n\nThen require the matching `live-smoke` round-trip: emit the new contract and\nconfirm the far side acts on it. A unit test asserting a **literal** (`[\"ask:0\",\n\"ask:1\"]`) proves only that the literal is unchanged — derive the expectation\nfrom the exported builder instead, so retiring the protocol fails the test that\npins it.\n\nPR #527 renamed the Telegram button protocol `ask:\u003Cindex>` → `reply:v1:\u003C0-7>`\nwith every gate green. Personal Agent heartbeat cards kept emitting the old\npayload for a week; taps did nothing, and a test pinning the literal stayed green\nthroughout.\n\n## 2. Process lane\n\nDo not build packages, run live smoke, inspect the fleet, or run release checks\nfor a docs\u002Fskills\u002Fprocess-only diff. Select tests that cover the edited contract.\nFor the engineering-skill surface, use:\n\n```bash\npnpm exec vitest run \\\n  scripts\u002F__tests__\u002Fcheck-codex-discoverability.test.mjs \\\n  scripts\u002F__tests__\u002Frepo-hygiene-skills.test.mjs \\\n  scripts\u002Frelease\u002F__tests__\u002Fpackage-count-drift.test.mjs\npnpm run check:codex-discoverability\npnpm run check:oss-hygiene\ngit diff --check\n```\n\nRun the website build only when `docs\u002F`, `website\u002F`, or published-site inputs\nchanged; use `docs-sync` to choose that surface.\n\n## 3. Package lane\n\nWhile iterating on package X:\n\n```bash\npnpm --filter @mono-agent\u002F\u003CX>... run build\npnpm --filter @mono-agent\u002F\u003CX> test\npnpm --filter @mono-agent\u002F\u003CX> run typecheck\n```\n\nAdd focused dependent-package checks when the changed public behavior crosses a\npackage boundary. Before merge, rely on one successful hosted CI run for the\nexact head SHA. If Actions is unavailable, run `pnpm run verify:all` once under\na supported Node runtime and record why CI could not provide the broad gate.\n\n## 4. High-risk lane\n\nRun the focused package checks above, then:\n\n```bash\npnpm run verify:all\n```\n\nRequire hosted CI for the exact head SHA and select exactly one scenario from\n`live-smoke` that exercises the risky boundary. Add a second scenario only when\nthe diff independently changes a second live surface.\n\nProvider-backed smoke is reserved for provider behavior. UI, worker transport,\nand lifecycle changes should use their narrower local scenarios instead of paid\nmodel calls.\n\n## Worktree dist rule\n\nCross-package imports resolve through built `dist\u002F`. After editing package A,\nrebuild A before testing or typechecking a dependent package. A process-only\nchange does not exercise package resolution and needs no dist baseline. See\n`worktree-feature` for diff-aware worktree setup.\n\n## Failure handling\n\n- Diagnose the first stable failure; do not rerun an unchanged failing broad gate.\n- When a failure may pre-exist, compare against the exact base SHA in a detached worktree.\n- A failure that passes on rerun at the same SHA is a flake: route it with `flaky-tests` (fix in scope, issue otherwise) instead of only rerunning.\n- If hosted CI fails immediately for an account or billing condition, stop polling and report that external blocker.\n- Never replace a failed stated check with a different check without naming the substitution.\n\n## Report format\n\nState the lane, exact commands, SHA, and pass\u002Ffail result. For failures, name the\nfailing test or gate. A green claim without command evidence is not complete.\n",{"name":5,"description":7},"main","fd0e4fa8a7f708a6e2a9c9546a5d2ca04956b06b","skills\u002Fverify-green",1,"https:\u002F\u002Fgithub.com\u002Frobertsreberski\u002Fmono-agent","https:\u002F\u002Fmono-agent.dev","https:\u002F\u002Fgithub.com\u002Frobertsreberski","linkedin.com\u002Fin\u002Frobert-sreberski","Sr. Growth Engineer @ Jetpack Marketing","2026-11-15 04:01:38",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":46},[36,39,42],{"tool":37,"command":38},"npx","npx skills add robertsreberski\u002Fmono-agent --skill verify-green",{"tool":40,"command":41},"bunx","bunx skills add robertsreberski\u002Fmono-agent --skill verify-green",{"tool":43,"command":44},"pnpm","pnpm dlx skills add robertsreberski\u002Fmono-agent --skill verify-green","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Frobertsreberski\u002Fmono-agent\u002Fblob\u002Fmain\u002Fskills\u002Fverify-green\u002FSKILL.md",true,[48],{"path":49,"resourceType":50,"resourceTypeDesc":51,"sizeBytes":52,"sourceHtmlUrl":53},"agents\u002Fopenai.yaml","REFERENCE","参考文档",226,"https:\u002F\u002Fgithub.com\u002Frobertsreberski\u002Fmono-agent\u002Fblob\u002Fmain\u002Fskills\u002Fverify-green\u002Fagents\u002Fopenai.yaml",[55,62,69,76,83,90],{"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},"01M74Q1JPF65A401Q1E0XE06VZ","worktree-feature","Start isolated mono-agent work with diff-aware setup, keep only required dist fresh, and ship through a PR with safe cleanup. Use for tracked repository changes, multi-commit work, or before executing an implementation plan.","https:\u002F\u002Fgithub.com\u002Frobertsreberski\u002Fmono-agent\u002Fblob\u002Fmain\u002Fskills\u002Fworktree-feature\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},"01M74Q1BY5KYZX3390MBVQYJTT","repo-hygiene-gc","Explicit bulk branch\u002Fworktree garbage collection using API-bound deletion proof. Use when asked to clean up the repository or run a deliberate maintenance sweep; ordinary PR cleanup belongs to worktree-feature.","https:\u002F\u002Fgithub.com\u002Frobertsreberski\u002Fmono-agent\u002Fblob\u002Fmain\u002Fskills\u002Frepo-hygiene-gc\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},"01M74Q19BRFX77BH3H8FT7YQD2","release-lockstep","Cut and registry-verify a lockstep npm release of all @mono-agent packages. Use when asked to release, publish to npm, cut vX.Y.Z, or bump the version. A release never implies consumer deployment.","https:\u002F\u002Fgithub.com\u002Frobertsreberski\u002Fmono-agent\u002Fblob\u002Fmain\u002Fskills\u002Frelease-lockstep\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},"01M74Q15Q68EJN55ZEMHTKKX8G","pi-upstream-recon","Investigate the vendored pi packages' real API surface (pi-ai and pi-agent-core) before hand-rolling anything, and run pi version bumps safely. Use before implementing runtime\u002Fprovider\u002Fsession behavior, when pi behavior is surprising, or when asked to \"bump pi\".","https:\u002F\u002Fgithub.com\u002Frobertsreberski\u002Fmono-agent\u002Fblob\u002Fmain\u002Fskills\u002Fpi-upstream-recon\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},"01M74Q13CYXZP9MASJ7ANWQ725","ops-log-hygiene","Check bounded logs for explicitly restarted services, or run a full log-health audit when requested or when symptoms justify it. Use for crash-loop diagnosis, oversized logs, restart churn, or an explicit live-fleet log audit.","https:\u002F\u002Fgithub.com\u002Frobertsreberski\u002Fmono-agent\u002Fblob\u002Fmain\u002Fskills\u002Fops-log-hygiene\u002FSKILL.md",[],[],{"skillNo":91,"name":92,"nameZh":6,"description":93,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":94,"categories":95,"keywords":96,"publishedAt":19},"01M74Q0ZS7SRD7HDJM6R6SR834","new-package","Add a new @mono-agent package\u002Fadapter that passes the architecture gate on the first try (catalog entry, README sections, dependency categories, lockstep version). Use when creating any new package, or when check:architecture fails on catalog\u002FREADME\u002Fboundary errors.","https:\u002F\u002Fgithub.com\u002Frobertsreberski\u002Fmono-agent\u002Fblob\u002Fmain\u002Fskills\u002Fnew-package\u002FSKILL.md",[],[]]