[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M732CB1WF6Z935ZAXMFSN3JR":3,"skill-same:01M732CB1WF6Z935ZAXMFSN3JR":46},{"skillNo":4,"name":5,"nameZh":6,"description":6,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":15,"categories":16,"keywords":17,"publishedAt":18,"linkOnly":19,"markdownContent":20,"markdownBody":21,"frontMatter":22,"refName":23,"commitSha":24,"skillDirectory":25,"forksCount":26,"repositoryHtmlUrl":27,"repositoryHomepageUrl":6,"ownerHtmlUrl":28,"ownerBlogUrl":29,"ownerDescription":30,"updatedAt":31,"install":32,"resources":45},"01M732CB1WF6Z935ZAXMFSN3JR","swarm-implement",null,"djimit\u002Fdjimitflo","djimit","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F19958832?v=4","01M7327CH792DZPFZPYAEFPRXT","Dutch Dim","01M7327CHWPSXT9AWE4MPWCX9Q",2,"MIT","https:\u002F\u002Fgithub.com\u002Fdjimit\u002Fdjimitflo\u002Fblob\u002Fmain\u002F.opencode\u002Fskills\u002Fswarm-implement\u002FSKILL.md",[],[],"2026-11-11 15:11:14",false,"---\nname: swarm-implement\ndescription: Execute complex implementation work with a swarm-like workflow: parallel exploration, scoped planning, objective validation, mandatory independent implementation review for changed work, and final critic approval. Use for feature work, bug fixes, refactors, and multi-file changes.\ndisable-model-invocation: true\n---\n\n# \u002Fswarm-implement\n\nUse this skill for implementation work when you want a fast, high-quality swarm\nworkflow rather than a single-threaded assistant.\n\n## Purpose\n\nComplete real coding tasks while preserving speed and adding swarm-style quality\ndiscipline.\n\n## Core operating model\n\nUse this execution ladder:\n\n1. Explore in parallel.\n2. Build a scoped plan.\n3. Implement in small, coherent units.\n4. Run objective validation.\n5. For any worktree edit, use independent reviewer validation on the latest diff\n   and evidence.\n6. For any worktree edit, use a separate final critic after reviewer approval.\n7. Synthesize and report what changed, what was verified, and what remains risky.\n\n## Command Namespace\n\nSwarm commands always use `\u002Fswarm \u003Csubcommand>`. Never invoke bare subcommand\nnames that collide with host commands such as `\u002Fplan`, `\u002Freset`, `\u002Fcheckpoint`,\nor `\u002Fstatus`.\n\n## High-risk work\n\nAlways use the deeper validation path for auth, permissions, payments,\ndestructive actions, dependency changes, public APIs, schemas, migrations,\nconcurrency, queues, retries, state machines, caching, file access, subprocesses,\nparsing, secrets, security-sensitive logic, and large cross-file refactors with\ncorrectness risk.\n\n## Recommended workflow\n\n### Phase 0 - Establish scope\n\nDetermine the exact task scope first:\n\n- what changed or needs to change,\n- what files are likely involved,\n- what success looks like,\n- what must not be broken,\n- what verification is required.\n\nIf the task is unclear, ask targeted questions or create a short written plan\nbefore coding.\n\n### Phase 0a - Parallel work check\n\nIf this project has the OpenCode-generated\n[`parallel-work-check` skill](..\u002Fgenerated\u002Fparallel-work-check\u002FSKILL.md), follow\nits full protocol. Otherwise, before starting implementation on an existing\nbranch:\n\n1. Fetch remote state and compare with local (`git fetch` plus HEAD hashes).\n2. If parallel swarm work is detected on the target branch, read the new commits,\n   decide whether to integrate, supersede, or proceed, and document the decision.\n3. Prefer the parallel work unless you can clearly articulate why your approach\n   is better.\n\n### Phase 0b - PR branch checkout pre-flight\n\nWhen implementation or review-scoping work depends on explorer agents reading a\nPR branch or commit range, complete this before Phase 1 explorer dispatch:\n\n1. Verify the working tree is clean with `git status --porcelain`. If\n   uncommitted changes exist, stash them or abort the checkout to prevent data\n   loss.\n2. Fetch and check out the PR head branch locally. Explorer agents read files\n   from the working tree (`Read`\u002F`Glob`\u002F`Grep`), not from git history, so a stale\n   checkout makes them inspect the base branch.\n3. Pass the exact commit range (`base_ref..head_ref`) in every explorer\n   delegation so agents have revision context for targeted `git show`\n   inspection.\n\n### Phase 1 - Parallel exploration\n\nLaunch parallel subagents for disjoint investigation tasks such as repository\nmapping, locating existing patterns, finding tests\u002Fcontracts, side-effect\nanalysis, and dependency or migration checks. Keep the main context focused.\n\n### Phase 2 - Plan\n\nCreate a concrete implementation plan before editing for any non-trivial task.\nInclude files to change, intended behavior, risks, validation commands, and\nwhether reviewer and critic passes are required.\n\n### Phase 3 - Implement in scoped units\n\nImplement in coherent, reviewable chunks. Follow existing repository patterns.\n\n**`declare_scope` discipline.** Before every coder or test-engineer delegation (and before every retry of one), call `declare_scope({ taskId, files })` with the exact file list the delegated agent is allowed to modify — including generated\u002Flockfile paths the change will produce (e.g. `dist\u002F*`, `package-lock.json`, `bun.lock`). Scope is enforced at the Edit\u002FWrite\u002FPatch tool layer only — bash-based writes (`sed -i`, `echo >`, `cat > \u003C\u003CHEREDOC`, heredoc-to-file redirects) bypass the check (see issue #520), so do not rely on `declare_scope` to bound shell-redirect writes. If the file list is not 100% obvious from the prompt, declare the containing directories instead. Repeat the declaration before every retry, even when files appear unchanged — scope state is per-attempt, not per-task.\n\nIf the project exposes `sast_scan` with `capture_baseline`, capture the\nphase-scoped SAST baseline before first coder delegation so later scans fail only\non new findings rather than pre-existing infrastructure noise.\n\n### Phase 4 - Objective validation\n\nRun the strongest objective checks available for the task: tests, lint,\ntypecheck, build, targeted repro scripts, and local runtime verification where\nrelevant. If you cannot verify it, do not claim it is done.\n\n### Phase 5 - Independent reviewer validation\n\nUse an independent reviewer subagent when the task edits code, tests, docs,\npackage metadata, release notes, or skill files. The reviewer must inspect the\nactual current diff and validation evidence after implementation. Any later edit\ninvalidates approval and requires re-review.\n\nReviewer responsibilities:\n\n- inspect the implementation with fresh context,\n- look for correctness bugs, edge cases, regressions, claim-vs-actual\n  mismatches, and test blind spots,\n- default to disbelief until evidence supports the change,\n- classify issues as `CONFIRMED`, `DISPROVED`, `UNVERIFIED`, or `PRE_EXISTING`\n  when useful,\n- verify that at least one regression test in the diff has falsification\n  evidence when the task claims regression coverage: the fix was temporarily\n  removed or bypassed, the test failed for the expected reason, the fix was\n  restored, and the test passed again. If no regression test is applicable, the\n  reviewer must record why,\n- return `APPROVE`, `NEEDS_REVISION`, or `BLOCKED`.\n\n`NEEDS_REVISION` or `BLOCKED` blocks completion until fixed and re-reviewed.\n\n### Phase 6 - Critic challenge\n\nUse a critic subagent after independent reviewer approval for any task that edits\ncode, tests, docs, package metadata, release notes, or skill files. The critic\nmust challenge the reviewer-approved current diff and evidence. Any later edit\ninvalidates critic approval and requires another critic pass.\n\n### Phase 7 - Final synthesis\n\nBefore calling work complete, verify:\n\n- objective validation ran and results were recorded,\n- the independent reviewer approved the latest diff and evidence,\n- a separate critic approved after reviewer approval,\n- every `NEEDS_REVISION` or `BLOCKED` item was fixed and re-reviewed,\n- no edit occurred after the latest reviewer\u002Fcritic approval,\n- no behavior is unwired or deferred without explicit user instruction.\n\n## Adapter notes\n\nRuntime-specific adapter skills in `.claude\u002Fskills\u002Fswarm-implement\u002F` and\n`.agents\u002Fskills\u002Fswarm-implement\u002F` must stay thin and delegate to this canonical\n`.opencode` workflow.\n","# \u002Fswarm-implement\n\nUse this skill for implementation work when you want a fast, high-quality swarm\nworkflow rather than a single-threaded assistant.\n\n## Purpose\n\nComplete real coding tasks while preserving speed and adding swarm-style quality\ndiscipline.\n\n## Core operating model\n\nUse this execution ladder:\n\n1. Explore in parallel.\n2. Build a scoped plan.\n3. Implement in small, coherent units.\n4. Run objective validation.\n5. For any worktree edit, use independent reviewer validation on the latest diff\n   and evidence.\n6. For any worktree edit, use a separate final critic after reviewer approval.\n7. Synthesize and report what changed, what was verified, and what remains risky.\n\n## Command Namespace\n\nSwarm commands always use `\u002Fswarm \u003Csubcommand>`. Never invoke bare subcommand\nnames that collide with host commands such as `\u002Fplan`, `\u002Freset`, `\u002Fcheckpoint`,\nor `\u002Fstatus`.\n\n## High-risk work\n\nAlways use the deeper validation path for auth, permissions, payments,\ndestructive actions, dependency changes, public APIs, schemas, migrations,\nconcurrency, queues, retries, state machines, caching, file access, subprocesses,\nparsing, secrets, security-sensitive logic, and large cross-file refactors with\ncorrectness risk.\n\n## Recommended workflow\n\n### Phase 0 - Establish scope\n\nDetermine the exact task scope first:\n\n- what changed or needs to change,\n- what files are likely involved,\n- what success looks like,\n- what must not be broken,\n- what verification is required.\n\nIf the task is unclear, ask targeted questions or create a short written plan\nbefore coding.\n\n### Phase 0a - Parallel work check\n\nIf this project has the OpenCode-generated\n[`parallel-work-check` skill](..\u002Fgenerated\u002Fparallel-work-check\u002FSKILL.md), follow\nits full protocol. Otherwise, before starting implementation on an existing\nbranch:\n\n1. Fetch remote state and compare with local (`git fetch` plus HEAD hashes).\n2. If parallel swarm work is detected on the target branch, read the new commits,\n   decide whether to integrate, supersede, or proceed, and document the decision.\n3. Prefer the parallel work unless you can clearly articulate why your approach\n   is better.\n\n### Phase 0b - PR branch checkout pre-flight\n\nWhen implementation or review-scoping work depends on explorer agents reading a\nPR branch or commit range, complete this before Phase 1 explorer dispatch:\n\n1. Verify the working tree is clean with `git status --porcelain`. If\n   uncommitted changes exist, stash them or abort the checkout to prevent data\n   loss.\n2. Fetch and check out the PR head branch locally. Explorer agents read files\n   from the working tree (`Read`\u002F`Glob`\u002F`Grep`), not from git history, so a stale\n   checkout makes them inspect the base branch.\n3. Pass the exact commit range (`base_ref..head_ref`) in every explorer\n   delegation so agents have revision context for targeted `git show`\n   inspection.\n\n### Phase 1 - Parallel exploration\n\nLaunch parallel subagents for disjoint investigation tasks such as repository\nmapping, locating existing patterns, finding tests\u002Fcontracts, side-effect\nanalysis, and dependency or migration checks. Keep the main context focused.\n\n### Phase 2 - Plan\n\nCreate a concrete implementation plan before editing for any non-trivial task.\nInclude files to change, intended behavior, risks, validation commands, and\nwhether reviewer and critic passes are required.\n\n### Phase 3 - Implement in scoped units\n\nImplement in coherent, reviewable chunks. Follow existing repository patterns.\n\n**`declare_scope` discipline.** Before every coder or test-engineer delegation (and before every retry of one), call `declare_scope({ taskId, files })` with the exact file list the delegated agent is allowed to modify — including generated\u002Flockfile paths the change will produce (e.g. `dist\u002F*`, `package-lock.json`, `bun.lock`). Scope is enforced at the Edit\u002FWrite\u002FPatch tool layer only — bash-based writes (`sed -i`, `echo >`, `cat > \u003C\u003CHEREDOC`, heredoc-to-file redirects) bypass the check (see issue #520), so do not rely on `declare_scope` to bound shell-redirect writes. If the file list is not 100% obvious from the prompt, declare the containing directories instead. Repeat the declaration before every retry, even when files appear unchanged — scope state is per-attempt, not per-task.\n\nIf the project exposes `sast_scan` with `capture_baseline`, capture the\nphase-scoped SAST baseline before first coder delegation so later scans fail only\non new findings rather than pre-existing infrastructure noise.\n\n### Phase 4 - Objective validation\n\nRun the strongest objective checks available for the task: tests, lint,\ntypecheck, build, targeted repro scripts, and local runtime verification where\nrelevant. If you cannot verify it, do not claim it is done.\n\n### Phase 5 - Independent reviewer validation\n\nUse an independent reviewer subagent when the task edits code, tests, docs,\npackage metadata, release notes, or skill files. The reviewer must inspect the\nactual current diff and validation evidence after implementation. Any later edit\ninvalidates approval and requires re-review.\n\nReviewer responsibilities:\n\n- inspect the implementation with fresh context,\n- look for correctness bugs, edge cases, regressions, claim-vs-actual\n  mismatches, and test blind spots,\n- default to disbelief until evidence supports the change,\n- classify issues as `CONFIRMED`, `DISPROVED`, `UNVERIFIED`, or `PRE_EXISTING`\n  when useful,\n- verify that at least one regression test in the diff has falsification\n  evidence when the task claims regression coverage: the fix was temporarily\n  removed or bypassed, the test failed for the expected reason, the fix was\n  restored, and the test passed again. If no regression test is applicable, the\n  reviewer must record why,\n- return `APPROVE`, `NEEDS_REVISION`, or `BLOCKED`.\n\n`NEEDS_REVISION` or `BLOCKED` blocks completion until fixed and re-reviewed.\n\n### Phase 6 - Critic challenge\n\nUse a critic subagent after independent reviewer approval for any task that edits\ncode, tests, docs, package metadata, release notes, or skill files. The critic\nmust challenge the reviewer-approved current diff and evidence. Any later edit\ninvalidates critic approval and requires another critic pass.\n\n### Phase 7 - Final synthesis\n\nBefore calling work complete, verify:\n\n- objective validation ran and results were recorded,\n- the independent reviewer approved the latest diff and evidence,\n- a separate critic approved after reviewer approval,\n- every `NEEDS_REVISION` or `BLOCKED` item was fixed and re-reviewed,\n- no edit occurred after the latest reviewer\u002Fcritic approval,\n- no behavior is unwired or deferred without explicit user instruction.\n\n## Adapter notes\n\nRuntime-specific adapter skills in `.claude\u002Fskills\u002Fswarm-implement\u002F` and\n`.agents\u002Fskills\u002Fswarm-implement\u002F` must stay thin and delegate to this canonical\n`.opencode` workflow.\n",{},"main","1667541871a3a917f9e544e5584ef06a544a119d",".opencode\u002Fskills\u002Fswarm-implement",0,"https:\u002F\u002Fgithub.com\u002Fdjimit\u002Fdjimitflo","https:\u002F\u002Fgithub.com\u002Fdjimit","https:\u002F\u002Fdjimit.nl\u002F","About me\r\n\r\nI’m Dennis a strategic IT advisor specializing in cloud, cybersecurity, privacy, AI, and enterprise architecture. ","2026-11-15 04:42:36",{"source":7,"skillName":5,"commands":33,"prompt":43,"zipAvailable":44},[34,37,40],{"tool":35,"command":36},"npx","npx skills add djimit\u002Fdjimitflo --skill swarm-implement",{"tool":38,"command":39},"bunx","bunx skills add djimit\u002Fdjimitflo --skill swarm-implement",{"tool":41,"command":42},"pnpm","pnpm dlx skills add djimit\u002Fdjimitflo --skill swarm-implement","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fdjimit\u002Fdjimitflo\u002Fblob\u002Fmain\u002F.opencode\u002Fskills\u002Fswarm-implement\u002FSKILL.md",true,[],[47,54,61,68,75,82],{"skillNo":48,"name":49,"nameZh":6,"description":50,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":51,"categories":52,"keywords":53,"publishedAt":18},"01M732DRW7MHQP9VTJ5JSBMWC3","provider-step-usage-v1-0-0","Sum measured OpenCode step_finish token totals with replay deduplication and fail-closed validation.","https:\u002F\u002Fgithub.com\u002Fdjimit\u002Fdjimitflo\u002Fblob\u002Fmain\u002Fskills\u002Fprovider-step-usage-v1-0-0\u002FSKILL.md",[],[],{"skillNo":55,"name":56,"nameZh":6,"description":57,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":58,"categories":59,"keywords":60,"publishedAt":18},"01M732DKXHWZ1P99ZDNA4PX7BV","writing-tests","Guidelines for writing, organizing, and maintaining tests in the opencode-swarm repository. Covers framework rules (bun:test), mock isolation, CI pipeline structure, file placement, and anti-patterns that break cross-platform CI. Load this skill before writing or modifying any test file.","https:\u002F\u002Fgithub.com\u002Fdjimit\u002Fdjimitflo\u002Fblob\u002Fmain\u002F.opencode\u002Fskills\u002Fwriting-tests\u002FSKILL.md",[],[],{"skillNo":62,"name":63,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":65,"categories":66,"keywords":67,"publishedAt":18},"01M732DCEG1R9S45A68T5WR2T7","worktree-retry-cleanup","Protocol for cleaning parallel-coder worktree lanes before retry. Triggered before re-dispatching any task that already has a lane (completed, denied, cancelled, or failed).","https:\u002F\u002Fgithub.com\u002Fdjimit\u002Fdjimitflo\u002Fblob\u002Fmain\u002F.opencode\u002Fskills\u002Fworktree-retry-cleanup\u002FSKILL.md",[],[],{"skillNo":69,"name":70,"nameZh":6,"description":71,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":72,"categories":73,"keywords":74,"publishedAt":18},"01M732D8KYWE5RHJ1AA76F70K7","swarm","Cross-agent swarm-mode behavior model — a higher-rigor workflow using parallel investigation, independent reviewer validation, and critic challenge, plus the mandatory implementation closeout gate. Runtime adapters (.claude, .agents) add execution-specific notes and command wiring.","https:\u002F\u002Fgithub.com\u002Fdjimit\u002Fdjimitflo\u002Fblob\u002Fmain\u002F.opencode\u002Fskills\u002Fswarm\u002FSKILL.md",[],[],{"skillNo":76,"name":77,"nameZh":6,"description":78,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":79,"categories":80,"keywords":81,"publishedAt":18},"01M732D232XCMYP3XYKEAQC3YT","swarm-pr-subscribe","Monitor a pull request after creation and act autonomously on pushed PR activity. Use when subscribing to a PR after opening it, when asked to watch, babysit, or autofix a PR until merge, or when a \u003Cpr-activity> wake message or [pr-monitor:...] advisory arrives for a subscribed PR. Owns event triage (fix \u002F ask \u002F skip), bounded-retry escalation, and terminal-state cleanup.","https:\u002F\u002Fgithub.com\u002Fdjimit\u002Fdjimitflo\u002Fblob\u002Fmain\u002F.opencode\u002Fskills\u002Fswarm-pr-subscribe\u002FSKILL.md",[],[],{"skillNo":83,"name":84,"nameZh":6,"description":85,"descriptionZh":6,"repositoryFullName":7,"ownerLogin":8,"ownerAvatarUrl":9,"ownerAccountNo":10,"ownerDisplayName":11,"repositoryNo":12,"starsCount":13,"licenseSpdxId":14,"licenseFileUrl":6,"sourceHtmlUrl":86,"categories":87,"keywords":88,"publishedAt":18},"01M732CVKJ4QEN83DQ30EYG75A","swarm-pr-review","Run a graph-guided, tool-augmented Swarm PR review using context packing, parallel exploration, triggered plugin micro-lanes, independent reviewer validation, critic challenge, and metrics writeback. Use for deep pull request review with low false-positive tolerance and high recall.","https:\u002F\u002Fgithub.com\u002Fdjimit\u002Fdjimitflo\u002Fblob\u002Fmain\u002F.opencode\u002Fskills\u002Fswarm-pr-review\u002FSKILL.md",[],[]]