[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M77H1QCSYZMBHQVYSGEFSK31":3,"skill-same:01M77H1QCSYZMBHQVYSGEFSK31":52},{"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":30,"commitSha":31,"skillDirectory":32,"forksCount":33,"repositoryHtmlUrl":34,"repositoryHomepageUrl":35,"ownerHtmlUrl":36,"ownerBlogUrl":37,"ownerDescription":6,"updatedAt":38,"install":39,"resources":51},"01M77H1QCSYZMBHQVYSGEFSK31","issue-queue",null,"One repeatable, evidence-backed pass that prioritizes and resolves a bounded batch of backlog issues worst-first using specialized subagents. Inspects the queue via `fak issue-orchestrator` and `fak dispatch order`, partitions into ready-leaves, triage, and subdivide cohorts, arbitrates lane leases via `dos arbitrate`, dispatches isolated workers, independently witnesses reproduction tests, re-measures burndown with `--compare`, and commits with `(fak \u003Cleaf>)`. Use when managing the issue queue, picking the next bounded batch of issues, or advancing issue burndown.","anthony-chaudhary\u002Ffak","anthony-chaudhary","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F18080164?v=4","01M77GSKCAEKV3JS07J40B94DY","Anthony Chaudhary","01M77GSKD26Z7F4FQQD4VA30VX",48,"Apache-2.0","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fissue-queue\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: issue-queue\ndescription: One repeatable, evidence-backed pass that prioritizes and resolves a bounded batch of backlog issues worst-first using specialized subagents. Inspects the queue via `fak issue-orchestrator` and `fak dispatch order`, partitions into ready-leaves, triage, and subdivide cohorts, arbitrates lane leases via `dos arbitrate`, dispatches isolated workers, independently witnesses reproduction tests, re-measures burndown with `--compare`, and commits with `(fak \u003Cleaf>)`. Use when managing the issue queue, picking the next bounded batch of issues, or advancing issue burndown.\ndisable-model-invocation: false\nuser-invocable: true\nallowed-tools: Read, Bash, Write, Edit, Grep, Glob, Task\nargument-hint: \"[--lane \u003Cname>] [--top N] [--subdivide] [--triage] [--from-issues \u003Cpath>]  (no args = baseline + resolve top 1-3 ready issues)\"\nmetadata:\n  generated-by: fak project-assets sync\n  canonical: ..\u002F..\u002F..\u002F.claude\u002Fskills\u002Fissue-queue\u002FSKILL.md\n---\n\n# \u002Fissue-queue — manage, queue, and resolve backlog issues in bounded batches\n\n> **What this does.** Every open issue in the repository represents a pending deliverable, bug,\n> or feature. When left unmanaged, the issue backlog rots into sprawling, un-prioritized,\n> collision-prone work.\n>\n> While `\u002Fissue-orchestrator` coordinates large-scale multi-wave campaigns across many workers,\n> `\u002Fissue-queue` executes the **atomic, bounded issue management and resolution pass**:\n> baseline and inspect the issue queue, partition issues into actionable cohorts (ready dispatchable\n> leaves, epics needing decomposition, and scope triage), pick a bounded batch of 1–3 high-priority\n> units, arbitrate lane leases (`dos arbitrate`), delegate implementation to a coordinated team of\n> specialized subagents (`researcher`, `worker`, `cross-validator`, `issue-auditor`), independently\n> witness the reproduction test and fix, prove burndown with `--compare`, and commit by explicit path.\n\nThe shape: **baseline issue queue (`fak issue-queue --json` \u002F `fak dispatch order`) → partition into ready\u002Fsubdivide\u002Ftriage cohorts → research & explore candidates (parallel `researcher` \u002F `explore` subagents) → arbitrate leases (`dos arbitrate`) → parallel implementation wave (`worker` subagents via `task`) → adversarial verification & gap audit (`cross-validator` & `issue-auditor` subagents) → prove queue burndown (`--compare`) → commit cleanly by explicit path with `(#N)` and `(fak \u003Clane>)` → release lease.**\n\n---\n\n## The Rule of Bounded Batches\n\nNever attempt a repository-wide sweep in one pass. Enforce strict scoping:\n\n1. **Batch Scaling (4–8 issues)**: Scale the batch to 4–8 issues whenever candidate issues map to pairwise tree-disjoint lanes (`dos.toml`). Focus on atomic leaves with verified independent packages.\n2. **Shift-left proof by default**: Every bug fix must ship with a reproduction unit test that fails before the fix and passes after. Features must ship with concrete unit\u002Fcontract tests. Never accept mocks or self-report claims.\n3. **Queue taxonomy discipline**:\n   - **Ready Leaves**: Atomic S0\u002FS1 tasks with a single witness and bounded package tree (`internal\u002F\u003Clane>\u002F`).\n   - **Subdivide Queue**: Epics or multi-subsystem issues (>15 expected steps or touching 3+ packages) routed to decomposition before dispatch.\n   - **Triage Queue**: Issues with missing acceptance criteria, ambiguous scope, or missing likely files routed to repair before dispatch.\n   - **Cooling \u002F Held Queue**: Issues under backoff cooldown or held by active lane leases in `.dos\u002Flane-journal.jsonl`.\n4. **Collision safety**: Pre-dispatch lane arbitration via `dos arbitrate` is mandatory before touching files.\n5. **Coordinator stays clean via specialized subagents**: Substantive investigation, code edits, and verification are delegated to isolated subagents via the `task` tool (Gemini 3.8 Flash high `variant: high`). The coordinator verifies, witnesses, and commits.\n6. **Non-forgeable witness**: Issues are closed only by git ancestry on trunk (`Fixes #N` with `(fak \u003Clane>)`), verified by `dos commit-audit` and `dos verify`.\n\n---\n\n## Step 1 — Measure Baseline & Inventory Queue\n\nCapture the open backlog baseline and inspect queue cohorts using allocated scratch storage (`fak tree-doctor --scratch-dir issue-queue`):\n\n```bash\n# Allocate scratch directory for queue artifacts:\nfak tree-doctor --scratch-dir issue-queue\n\n# Snapshot current issue queue state:\nfak issue-queue --json > _scratch\u002Fissue-queue\u002Fbaseline.json\n\n# View top prioritized issues across lanes:\nfak issue-queue --top 5\n```\n\nIf querying live GitHub issues:\n```bash\ngh issue list --state open --json number,title,body,labels > _scratch\u002Fissue-queue\u002Flive-issues.json\nfak issue-queue --from-issues _scratch\u002Fissue-queue\u002Flive-issues.json --json > _scratch\u002Fissue-queue\u002Fbaseline.json\nfak issue-queue --from-issues _scratch\u002Fissue-queue\u002Flive-issues.json --top 5\n```\n\nNote the headline queue metrics:\n- **Total Issues Evaluated**: total backlog size.\n- **Dispatchable Ready Leaves**: atomic S0\u002FS1 issues ready for immediate dispatch.\n- **Subdivide Epics**: oversized issues requiring decomposition.\n- **Triage Issues**: issues requiring scope or acceptance criteria repair.\n- **Held Excluded**: issues currently held by active leases in `.dos\u002Flane-journal.jsonl`.\n\n---\n\n## Step 2 — Parallel Scoping & Pre-Dispatch Arbitration\n\n### 1. Select Bounded Target Batch (1–3 Issues)\nChoose 1–3 target issues based on leverage:\n- **Option A: High Priority \u002F High Centrality Hotspots (`P0`\u002F`P1`, `Core` centrality)**:\n  Issues in core leaves (`gateway`, `engine`, `model`, `compute`) with high inbound impact.\n- **Option B: Quick Leaf Wins (Highest Velocity)**:\n  Self-contained bug fixes or feature additions sitting at 2–4 expected steps with bounded write surfaces.\n\n### 2. Pre-Dispatch Research & Path Mapping (Parallel Subagents)\nLaunch parallel specialized discovery subagents in a single turn to research candidate context and verify disjoint boundaries without polluting coordinator context:\n\n```json\n[\n  {\n    \"tool\": \"task\",\n    \"parameters\": {\n      \"subagent_type\": \"researcher\",\n      \"description\": \"Research #1024 gateway timeout context\",\n      \"prompt\": \"Investigate issue #1024: Fix gateway streaming timeout.\\nAnalyze relevant files under internal\u002Fgateway\u002F, recent commits, and existing test patterns.\\nReturn a compact receipt: (1) root cause hypothesis, (2) exact files to edit, (3) proposed reproduction test strategy.\"\n    }\n  },\n  {\n    \"tool\": \"task\",\n    \"parameters\": {\n      \"subagent_type\": \"explore\",\n      \"description\": \"Explore model KV cache paths\",\n      \"prompt\": \"Inspect internal\u002Fmodel\u002F for KV cache allocation and recycling paths for issue #1035.\\nVerify that its write tree is strictly confined to internal\u002Fmodel\u002F and shares no files with internal\u002Fgateway\u002F.\"\n    }\n  }\n]\n```\n\n*(If an issue involves deep architectural or concurrency invariants, launch a `deep-reason` subagent for lock ordering and invariant analysis).*\n\n### 3. Verify Smallness & Contract\n```bash\n# Verify smallness \u002F atomicity (single deliverable, witness == 1):\nfak dispatch issue-smallness-lint --issue \u003Cnumber> --json\n\n# Verify issue contract and structure:\nfak-dev issue contract --from-issues _scratch\u002Fissue-queue\u002Flive-issues.json --json\n```\n\n### 4. Arbitrate Lane Leases\nVerify lane lease availability before dispatch:\n```bash\ndos arbitrate --workspace . --lane \u003Clane> --kind keyword --mode exclusive\n```\n*(Or invoke the `dos_arbitrate` MCP tool).*\n\n- **Outcome `acquire`**: Admitted. Proceed to dispatch.\n- **Outcome `refuse`**: Lane is held by a peer. Pick another candidate from the queue or wait out the lease.\n\n---\n\n## Step 3 — Parallel Implementation Worker Wave\n\nLaunch all selected issues' implementation workers in parallel in a single turn using multiple `task` tool calls with `subagent_type=\"worker\"` (using Gemini 3.8 Flash high `variant: high`). Each worker operates under strict package fences:\n\n```json\n[\n  {\n    \"tool\": \"task\",\n    \"parameters\": {\n      \"subagent_type\": \"worker\",\n      \"description\": \"Implement fix for #1024\",\n      \"prompt\": \"You are an isolated worker resolving issue #1024: Fix gateway streaming timeout.\\nLane: gateway\\nBoundaries: Edit ONLY files inside internal\u002Fgateway\u002F. NEVER touch go.mod, go.sum, dos.toml, or root files.\\nDeliverable:\\n  1. Write an atomic reproduction test in internal\u002Fgateway\u002Fstream_timeout_test.go demonstrating the defect.\\n  2. Implement the minimal fix in internal\u002Fgateway\u002F.\\n  3. Verify package tests pass.\\nVerification: Execute ONLY package-scoped tests: `go test -v .\u002Finternal\u002Fgateway` and `go vet .\u002Finternal\u002Fgateway`. NEVER run `go test .\u002F...`.\\nReturn Contract: Return a compact 3-line receipt: (1) files created\u002Fmodified, (2) package test results, (3) confirmation of fix.\"\n    }\n  },\n  {\n    \"tool\": \"task\",\n    \"parameters\": {\n      \"subagent_type\": \"worker\",\n      \"description\": \"Implement feature for #1035\",\n      \"prompt\": \"You are an isolated worker resolving issue #1035: Add model KV cache recycling.\\nLane: model\\nBoundaries: Edit ONLY files inside internal\u002Fmodel\u002F. NEVER touch go.mod, go.sum, dos.toml, or root files.\\nDeliverable:\\n  1. Write an atomic reproduction test in internal\u002Fmodel\u002Fkv_recycle_test.go.\\n  2. Implement the KV cache recycling path in internal\u002Fmodel\u002F.\\n  3. Verify package tests pass.\\nVerification: Execute ONLY package-scoped tests: `go test -v .\u002Finternal\u002Fmodel` and `go vet .\u002Finternal\u002Fmodel`. NEVER run `go test .\u002F...`.\\nReturn Contract: Return a compact 3-line receipt: (1) files created\u002Fmodified, (2) package test results, (3) confirmation of feature.\"\n    }\n  }\n]\n```\n\nEach worker executes the concrete implementation:\n1. **Reproduction Unit Test**: Focused test reproducing the defect or specifying the contract.\n2. **Minimal Package Implementation**: Code changes strictly confined to `internal\u002F\u003Clane>\u002F`.\n3. **Package Verification**: Scoped checks (`go vet .\u002Finternal\u002F\u003Clane>`, `go test -v .\u002Finternal\u002F\u003Clane>`).\n\n---\n\n## Step 4 — Parallel Adversarial Cross-Validation & Gap Audit\n\nThe coordinator does NOT trust worker narration (`dos-witness-claim`). Instead, launch parallel verification subagents to independently audit the landed work before commit:\n\n```json\n[\n  {\n    \"tool\": \"task\",\n    \"parameters\": {\n      \"subagent_type\": \"cross-validator\",\n      \"description\": \"Cross-validate #1024 and #1035\",\n      \"prompt\": \"You are an adversarial cross-validator.\\nTasks to verify: #1024 (internal\u002Fgateway) and #1035 (internal\u002Fmodel).\\nResponsibilities:\\n  1. Inspect `git diff internal\u002Fgateway` and `git diff internal\u002Fmodel` to verify that edits did not leak outside declared package boundaries.\\n  2. Execute on-device tests: `go test -v .\u002Finternal\u002Fgateway` and `go test -v .\u002Finternal\u002Fmodel`.\\n  3. Verify that reproduction tests genuinely assert the expected behavior and do not test trivial mocks.\\nReturn a structured DOS-style proof verdict: VERIFIED or REJECTED with concrete command witnesses.\"\n    }\n  },\n  {\n    \"tool\": \"task\",\n    \"parameters\": {\n      \"subagent_type\": \"issue-auditor\",\n      \"description\": \"Audit QA gaps and follow-ons\",\n      \"prompt\": \"You are an issue auditor inspecting the changes in internal\u002Fgateway and internal\u002Fmodel.\\nUncover unhandled edge cases, boundary conditions, or failure modes in the newly added code.\\nIf real follow-ons or gaps exist, draft structured GitHub issue tickets with clear reproduction steps and acceptance criteria.\"\n    }\n  }\n]\n```\n\nCoordinator independently verifies:\n1. **Toolchain check**: `go vet .\u002Finternal\u002F\u003Clane>` and `go test -v .\u002Finternal\u002F\u003Clane>`.\n2. **Diff boundary check**: `git diff --stat internal\u002F\u003Clane>`.\n3. **Audit review**: Incorporate cross-validator proof verdicts and any auditor follow-on tickets.\n\n---\n\n## Step 5 — Re-Measure and Prove Queue Burndown\n\nRe-evaluate the issue queue against the saved baseline to prove burndown:\n\n```bash\n# Re-snapshot issue state:\ngh issue list --state open --json number,title,body,labels > _scratch\u002Fissue-queue\u002Fupdated-issues.json\nfak issue-queue --from-issues _scratch\u002Fissue-queue\u002Fupdated-issues.json --compare _scratch\u002Fissue-queue\u002Fbaseline.json\n```\n\nConfirm:\n- Target issues are resolved and retired from the dispatchable count.\n- Closed count delta reflects the completed batch.\n- No new unclassified drift was introduced.\n\n---\n\n## Step 6 — Commit Cleanly by Explicit Path\n\nCommit each finished leaf independently on the trunk. Include a Conventional-Commits subject, signed-off DCO (`-s`), issue citation `(#N)`, and required `(fak \u003Cleaf>)` trailer:\n\n```bash\nfak sync reconcile --apply\nfak commit --path internal\u002F\u003ClaneA> -m \"fix(\u003ClaneA>): resolve \u003Csummary> (#1024) (fak \u003ClaneA>)\"\nfak commit --path internal\u002F\u003ClaneB> -m \"feat(\u003ClaneB>): add \u003Csummary> (#1035) (fak \u003ClaneB>)\"\nfak sync push\n```\n*(Fallback: `git commit -s -m \"...\" -- internal\u002F\u003Clane>` without `git add -A`).*\n\nVerify `git status` shows your lanes are committed cleanly and no peer WIP was staged.\n\n---\n\n## Step 7 — Release Lane Leases & Clean Scratch\n\nRelease all acquired lane leases:\n\n```bash\ndos lease-lane release --lane \u003ClaneA>\ndos lease-lane release --lane \u003ClaneB>\n```\n\nReap the campaign scratch directory:\n```bash\nfak tree-doctor --reap-scratch issue-queue --json\n```\n\nReport the one-line completion receipt to the operator with the checkable commit SHAs, issue numbers, and test witnesses.\n","# \u002Fissue-queue — manage, queue, and resolve backlog issues in bounded batches\n\n> **What this does.** Every open issue in the repository represents a pending deliverable, bug,\n> or feature. When left unmanaged, the issue backlog rots into sprawling, un-prioritized,\n> collision-prone work.\n>\n> While `\u002Fissue-orchestrator` coordinates large-scale multi-wave campaigns across many workers,\n> `\u002Fissue-queue` executes the **atomic, bounded issue management and resolution pass**:\n> baseline and inspect the issue queue, partition issues into actionable cohorts (ready dispatchable\n> leaves, epics needing decomposition, and scope triage), pick a bounded batch of 1–3 high-priority\n> units, arbitrate lane leases (`dos arbitrate`), delegate implementation to a coordinated team of\n> specialized subagents (`researcher`, `worker`, `cross-validator`, `issue-auditor`), independently\n> witness the reproduction test and fix, prove burndown with `--compare`, and commit by explicit path.\n\nThe shape: **baseline issue queue (`fak issue-queue --json` \u002F `fak dispatch order`) → partition into ready\u002Fsubdivide\u002Ftriage cohorts → research & explore candidates (parallel `researcher` \u002F `explore` subagents) → arbitrate leases (`dos arbitrate`) → parallel implementation wave (`worker` subagents via `task`) → adversarial verification & gap audit (`cross-validator` & `issue-auditor` subagents) → prove queue burndown (`--compare`) → commit cleanly by explicit path with `(#N)` and `(fak \u003Clane>)` → release lease.**\n\n---\n\n## The Rule of Bounded Batches\n\nNever attempt a repository-wide sweep in one pass. Enforce strict scoping:\n\n1. **Batch Scaling (4–8 issues)**: Scale the batch to 4–8 issues whenever candidate issues map to pairwise tree-disjoint lanes (`dos.toml`). Focus on atomic leaves with verified independent packages.\n2. **Shift-left proof by default**: Every bug fix must ship with a reproduction unit test that fails before the fix and passes after. Features must ship with concrete unit\u002Fcontract tests. Never accept mocks or self-report claims.\n3. **Queue taxonomy discipline**:\n   - **Ready Leaves**: Atomic S0\u002FS1 tasks with a single witness and bounded package tree (`internal\u002F\u003Clane>\u002F`).\n   - **Subdivide Queue**: Epics or multi-subsystem issues (>15 expected steps or touching 3+ packages) routed to decomposition before dispatch.\n   - **Triage Queue**: Issues with missing acceptance criteria, ambiguous scope, or missing likely files routed to repair before dispatch.\n   - **Cooling \u002F Held Queue**: Issues under backoff cooldown or held by active lane leases in `.dos\u002Flane-journal.jsonl`.\n4. **Collision safety**: Pre-dispatch lane arbitration via `dos arbitrate` is mandatory before touching files.\n5. **Coordinator stays clean via specialized subagents**: Substantive investigation, code edits, and verification are delegated to isolated subagents via the `task` tool (Gemini 3.8 Flash high `variant: high`). The coordinator verifies, witnesses, and commits.\n6. **Non-forgeable witness**: Issues are closed only by git ancestry on trunk (`Fixes #N` with `(fak \u003Clane>)`), verified by `dos commit-audit` and `dos verify`.\n\n---\n\n## Step 1 — Measure Baseline & Inventory Queue\n\nCapture the open backlog baseline and inspect queue cohorts using allocated scratch storage (`fak tree-doctor --scratch-dir issue-queue`):\n\n```bash\n# Allocate scratch directory for queue artifacts:\nfak tree-doctor --scratch-dir issue-queue\n\n# Snapshot current issue queue state:\nfak issue-queue --json > _scratch\u002Fissue-queue\u002Fbaseline.json\n\n# View top prioritized issues across lanes:\nfak issue-queue --top 5\n```\n\nIf querying live GitHub issues:\n```bash\ngh issue list --state open --json number,title,body,labels > _scratch\u002Fissue-queue\u002Flive-issues.json\nfak issue-queue --from-issues _scratch\u002Fissue-queue\u002Flive-issues.json --json > _scratch\u002Fissue-queue\u002Fbaseline.json\nfak issue-queue --from-issues _scratch\u002Fissue-queue\u002Flive-issues.json --top 5\n```\n\nNote the headline queue metrics:\n- **Total Issues Evaluated**: total backlog size.\n- **Dispatchable Ready Leaves**: atomic S0\u002FS1 issues ready for immediate dispatch.\n- **Subdivide Epics**: oversized issues requiring decomposition.\n- **Triage Issues**: issues requiring scope or acceptance criteria repair.\n- **Held Excluded**: issues currently held by active leases in `.dos\u002Flane-journal.jsonl`.\n\n---\n\n## Step 2 — Parallel Scoping & Pre-Dispatch Arbitration\n\n### 1. Select Bounded Target Batch (1–3 Issues)\nChoose 1–3 target issues based on leverage:\n- **Option A: High Priority \u002F High Centrality Hotspots (`P0`\u002F`P1`, `Core` centrality)**:\n  Issues in core leaves (`gateway`, `engine`, `model`, `compute`) with high inbound impact.\n- **Option B: Quick Leaf Wins (Highest Velocity)**:\n  Self-contained bug fixes or feature additions sitting at 2–4 expected steps with bounded write surfaces.\n\n### 2. Pre-Dispatch Research & Path Mapping (Parallel Subagents)\nLaunch parallel specialized discovery subagents in a single turn to research candidate context and verify disjoint boundaries without polluting coordinator context:\n\n```json\n[\n  {\n    \"tool\": \"task\",\n    \"parameters\": {\n      \"subagent_type\": \"researcher\",\n      \"description\": \"Research #1024 gateway timeout context\",\n      \"prompt\": \"Investigate issue #1024: Fix gateway streaming timeout.\\nAnalyze relevant files under internal\u002Fgateway\u002F, recent commits, and existing test patterns.\\nReturn a compact receipt: (1) root cause hypothesis, (2) exact files to edit, (3) proposed reproduction test strategy.\"\n    }\n  },\n  {\n    \"tool\": \"task\",\n    \"parameters\": {\n      \"subagent_type\": \"explore\",\n      \"description\": \"Explore model KV cache paths\",\n      \"prompt\": \"Inspect internal\u002Fmodel\u002F for KV cache allocation and recycling paths for issue #1035.\\nVerify that its write tree is strictly confined to internal\u002Fmodel\u002F and shares no files with internal\u002Fgateway\u002F.\"\n    }\n  }\n]\n```\n\n*(If an issue involves deep architectural or concurrency invariants, launch a `deep-reason` subagent for lock ordering and invariant analysis).*\n\n### 3. Verify Smallness & Contract\n```bash\n# Verify smallness \u002F atomicity (single deliverable, witness == 1):\nfak dispatch issue-smallness-lint --issue \u003Cnumber> --json\n\n# Verify issue contract and structure:\nfak-dev issue contract --from-issues _scratch\u002Fissue-queue\u002Flive-issues.json --json\n```\n\n### 4. Arbitrate Lane Leases\nVerify lane lease availability before dispatch:\n```bash\ndos arbitrate --workspace . --lane \u003Clane> --kind keyword --mode exclusive\n```\n*(Or invoke the `dos_arbitrate` MCP tool).*\n\n- **Outcome `acquire`**: Admitted. Proceed to dispatch.\n- **Outcome `refuse`**: Lane is held by a peer. Pick another candidate from the queue or wait out the lease.\n\n---\n\n## Step 3 — Parallel Implementation Worker Wave\n\nLaunch all selected issues' implementation workers in parallel in a single turn using multiple `task` tool calls with `subagent_type=\"worker\"` (using Gemini 3.8 Flash high `variant: high`). Each worker operates under strict package fences:\n\n```json\n[\n  {\n    \"tool\": \"task\",\n    \"parameters\": {\n      \"subagent_type\": \"worker\",\n      \"description\": \"Implement fix for #1024\",\n      \"prompt\": \"You are an isolated worker resolving issue #1024: Fix gateway streaming timeout.\\nLane: gateway\\nBoundaries: Edit ONLY files inside internal\u002Fgateway\u002F. NEVER touch go.mod, go.sum, dos.toml, or root files.\\nDeliverable:\\n  1. Write an atomic reproduction test in internal\u002Fgateway\u002Fstream_timeout_test.go demonstrating the defect.\\n  2. Implement the minimal fix in internal\u002Fgateway\u002F.\\n  3. Verify package tests pass.\\nVerification: Execute ONLY package-scoped tests: `go test -v .\u002Finternal\u002Fgateway` and `go vet .\u002Finternal\u002Fgateway`. NEVER run `go test .\u002F...`.\\nReturn Contract: Return a compact 3-line receipt: (1) files created\u002Fmodified, (2) package test results, (3) confirmation of fix.\"\n    }\n  },\n  {\n    \"tool\": \"task\",\n    \"parameters\": {\n      \"subagent_type\": \"worker\",\n      \"description\": \"Implement feature for #1035\",\n      \"prompt\": \"You are an isolated worker resolving issue #1035: Add model KV cache recycling.\\nLane: model\\nBoundaries: Edit ONLY files inside internal\u002Fmodel\u002F. NEVER touch go.mod, go.sum, dos.toml, or root files.\\nDeliverable:\\n  1. Write an atomic reproduction test in internal\u002Fmodel\u002Fkv_recycle_test.go.\\n  2. Implement the KV cache recycling path in internal\u002Fmodel\u002F.\\n  3. Verify package tests pass.\\nVerification: Execute ONLY package-scoped tests: `go test -v .\u002Finternal\u002Fmodel` and `go vet .\u002Finternal\u002Fmodel`. NEVER run `go test .\u002F...`.\\nReturn Contract: Return a compact 3-line receipt: (1) files created\u002Fmodified, (2) package test results, (3) confirmation of feature.\"\n    }\n  }\n]\n```\n\nEach worker executes the concrete implementation:\n1. **Reproduction Unit Test**: Focused test reproducing the defect or specifying the contract.\n2. **Minimal Package Implementation**: Code changes strictly confined to `internal\u002F\u003Clane>\u002F`.\n3. **Package Verification**: Scoped checks (`go vet .\u002Finternal\u002F\u003Clane>`, `go test -v .\u002Finternal\u002F\u003Clane>`).\n\n---\n\n## Step 4 — Parallel Adversarial Cross-Validation & Gap Audit\n\nThe coordinator does NOT trust worker narration (`dos-witness-claim`). Instead, launch parallel verification subagents to independently audit the landed work before commit:\n\n```json\n[\n  {\n    \"tool\": \"task\",\n    \"parameters\": {\n      \"subagent_type\": \"cross-validator\",\n      \"description\": \"Cross-validate #1024 and #1035\",\n      \"prompt\": \"You are an adversarial cross-validator.\\nTasks to verify: #1024 (internal\u002Fgateway) and #1035 (internal\u002Fmodel).\\nResponsibilities:\\n  1. Inspect `git diff internal\u002Fgateway` and `git diff internal\u002Fmodel` to verify that edits did not leak outside declared package boundaries.\\n  2. Execute on-device tests: `go test -v .\u002Finternal\u002Fgateway` and `go test -v .\u002Finternal\u002Fmodel`.\\n  3. Verify that reproduction tests genuinely assert the expected behavior and do not test trivial mocks.\\nReturn a structured DOS-style proof verdict: VERIFIED or REJECTED with concrete command witnesses.\"\n    }\n  },\n  {\n    \"tool\": \"task\",\n    \"parameters\": {\n      \"subagent_type\": \"issue-auditor\",\n      \"description\": \"Audit QA gaps and follow-ons\",\n      \"prompt\": \"You are an issue auditor inspecting the changes in internal\u002Fgateway and internal\u002Fmodel.\\nUncover unhandled edge cases, boundary conditions, or failure modes in the newly added code.\\nIf real follow-ons or gaps exist, draft structured GitHub issue tickets with clear reproduction steps and acceptance criteria.\"\n    }\n  }\n]\n```\n\nCoordinator independently verifies:\n1. **Toolchain check**: `go vet .\u002Finternal\u002F\u003Clane>` and `go test -v .\u002Finternal\u002F\u003Clane>`.\n2. **Diff boundary check**: `git diff --stat internal\u002F\u003Clane>`.\n3. **Audit review**: Incorporate cross-validator proof verdicts and any auditor follow-on tickets.\n\n---\n\n## Step 5 — Re-Measure and Prove Queue Burndown\n\nRe-evaluate the issue queue against the saved baseline to prove burndown:\n\n```bash\n# Re-snapshot issue state:\ngh issue list --state open --json number,title,body,labels > _scratch\u002Fissue-queue\u002Fupdated-issues.json\nfak issue-queue --from-issues _scratch\u002Fissue-queue\u002Fupdated-issues.json --compare _scratch\u002Fissue-queue\u002Fbaseline.json\n```\n\nConfirm:\n- Target issues are resolved and retired from the dispatchable count.\n- Closed count delta reflects the completed batch.\n- No new unclassified drift was introduced.\n\n---\n\n## Step 6 — Commit Cleanly by Explicit Path\n\nCommit each finished leaf independently on the trunk. Include a Conventional-Commits subject, signed-off DCO (`-s`), issue citation `(#N)`, and required `(fak \u003Cleaf>)` trailer:\n\n```bash\nfak sync reconcile --apply\nfak commit --path internal\u002F\u003ClaneA> -m \"fix(\u003ClaneA>): resolve \u003Csummary> (#1024) (fak \u003ClaneA>)\"\nfak commit --path internal\u002F\u003ClaneB> -m \"feat(\u003ClaneB>): add \u003Csummary> (#1035) (fak \u003ClaneB>)\"\nfak sync push\n```\n*(Fallback: `git commit -s -m \"...\" -- internal\u002F\u003Clane>` without `git add -A`).*\n\nVerify `git status` shows your lanes are committed cleanly and no peer WIP was staged.\n\n---\n\n## Step 7 — Release Lane Leases & Clean Scratch\n\nRelease all acquired lane leases:\n\n```bash\ndos lease-lane release --lane \u003ClaneA>\ndos lease-lane release --lane \u003ClaneB>\n```\n\nReap the campaign scratch directory:\n```bash\nfak tree-doctor --reap-scratch issue-queue --json\n```\n\nReport the one-line completion receipt to the operator with the checkable commit SHAs, issue numbers, and test witnesses.\n",{"name":5,"metadata":24,"description":7,"allowed-tools":27,"argument-hint":28,"user-invocable":29,"disable-model-invocation":20},{"canonical":25,"generated-by":26},"..\u002F..\u002F..\u002F.claude\u002Fskills\u002Fissue-queue\u002FSKILL.md","fak project-assets sync","Read, Bash, Write, Edit, Grep, Glob, Task","[--lane \u003Cname>] [--top N] [--subdivide] [--triage] [--from-issues \u003Cpath>]  (no args = baseline + resolve top 1-3 ready issues)",true,"main","f565b7b7db63e79649f3a02c08108132a86c5d05",".agents\u002Fskills\u002Fissue-queue",16,"https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak","https:\u002F\u002Fanthony-chaudhary.github.io\u002Ffak\u002Fshowcase.html","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary","https:\u002F\u002Flinkedin.com\u002Fin\u002Fanthonychaudhary","2026-11-15 03:06:00",{"source":8,"skillName":5,"commands":40,"prompt":50,"zipAvailable":29},[41,44,47],{"tool":42,"command":43},"npx","npx skills add anthony-chaudhary\u002Ffak --skill issue-queue",{"tool":45,"command":46},"bunx","bunx skills add anthony-chaudhary\u002Ffak --skill issue-queue",{"tool":48,"command":49},"pnpm","pnpm dlx skills add anthony-chaudhary\u002Ffak --skill issue-queue","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002F.agents\u002Fskills\u002Fissue-queue\u002FSKILL.md",[],[53,60,69,76,83,88],{"skillNo":54,"name":55,"nameZh":6,"description":56,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":57,"categories":58,"keywords":59,"publishedAt":19},"01M78V4059KRP2T88YR6Z05MJQ","serve-receipt","On-demand research protocol + deterministic receipt for serving a Hugging Face model on a concrete GPU target — fan out research lanes (model facts\u002Ffiles, engine support matrix, serving recipes, KV\u002Fstate memory math), compute the tiered feasibility verdict with `fak serve-receipt` (never narrated math), and emit a witnessed launch receipt. Use when asked for an \"initial config\", \"serving receipt\", or \"how do I serve \u003Cmodel> on \u003CGPUs>\".","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fserve-receipt\u002FSKILL.md",[],[],{"skillNo":61,"name":62,"nameZh":6,"description":63,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":64,"licenseFileUrl":65,"sourceHtmlUrl":66,"categories":67,"keywords":68,"publishedAt":19},"01M77HRGKQV2CJ2K6ZFN4YQ32K","caveman","Ultra-compressed communication mode. Cuts output tokens 65% (measured) by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says \"caveman mode\", \"talk like caveman\", \"use caveman\", \"less tokens\", \"be brief\", or invokes \u002Fcaveman. Also auto-triggers when token efficiency is requested.","MIT","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fsyspromptmmu\u002Fthird_party\u002Fcaveman\u002FLICENSE","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fsyspromptmmu\u002Fthird_party\u002Fcaveman\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},"01M77HRAY3Y3BE7APN5PVJF7NV","quality-report","Score the repo quality and emit a report card.","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fctxplans\u002Ftestdata\u002Frepo\u002F.claude\u002Fskills\u002Fquality-report\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},"01M77HR9807KHAX13DG9PW2K2Q","ctx-overlay","Compact the context window and prune the memory store when a session grows too long.","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fctxplans\u002Ftestdata\u002Frepo\u002F.claude\u002Fskills\u002Fctx-overlay\u002FSKILL.md",[],[],{"skillNo":84,"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":85,"categories":86,"keywords":87,"publishedAt":19},"01M77HR7XQ9X8F2CY9Z4F2ZEZ3","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fctxplanlint\u002Ftestdata\u002Frepo\u002F.claude\u002Fskills\u002Fquality-report\u002FSKILL.md",[],[],{"skillNo":89,"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":90,"categories":91,"keywords":92,"publishedAt":19},"01M77HR5BVKCTR1WNJGRBSDNX7","https:\u002F\u002Fgithub.com\u002Fanthony-chaudhary\u002Ffak\u002Fblob\u002Fmain\u002Finternal\u002Fctxplanlint\u002Ftestdata\u002Frepo\u002F.claude\u002Fskills\u002Fctx-overlay\u002FSKILL.md",[],[]]