How do I install the Branch And Task Management skill?

Run npx skillmds@latest add synaptiai/branch-and-task-management in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Branch And Task Management safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Branch And Task Management?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Branch And Task Management free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Branch And Task Management?

synaptiai (@synaptiai) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Branch And Task Management by synaptiai

Name: Branch And Task Management
Author: synaptiai

Branch And Task Management

Create feature branches with naming conventions, load full issue context and impact analysis, and decompose acceptance criteria into atomic parallel tasks with dependencies. Use when starting work on a GitHub issue. This skill MUST be consulted because starting code without context causes misaligned implementations and wasted effort.

synaptiai 8e3b65a 3.4 KB Updated Sep 17, 2026

File contents

Branch and Task Management

Contract

Iron law: no code before context — read the issue body and comments, load history, understand scope, then create the branch. Invoked by /flow:start (Phase 1 EXPLORE through Phase 2 PLAN) and by the implementation-planner agent for task decomposition. Returns: a feature branch named per conventions.branchPatterns checked out from the default branch, one TaskCreate task per acceptance criterion with dependencies set, and an initialized decision journal at {JOURNAL_DIR}/issue-{N}.md. Permitted skips: none — an issue without acceptance criteria is not decomposed; it goes back to the user (or Skill(issue-crafting)) first.

Pre-Conditions

Before creating a branch, confirm: the issue exists and is OPEN; it has acceptance criteria (otherwise ask the user or create them); no existing branch already addresses it; you have read the full body AND comments.

Branch Creation

DEFAULT_BRANCH=$(gh repo view --json defaultBranchRef --jq '.defaultBranchRef.name' 2>/dev/null || echo "main")
git fetch origin "$DEFAULT_BRANCH"
git checkout -b "feature/issue-{N}-{desc}" "origin/$DEFAULT_BRANCH"

Patterns from settings.json → conventions.branchPatterns: feature/issue-{N}-{desc}, fix/issue-{N}-{desc}, docs/issue-{N}-{desc}. {desc} is 3-5 kebab-case words.

Issue Context Loading

Fetch in parallel: gh issue view $N --json title,body,labels,assignees,milestone and gh issue view $N --comments. Extract: title; acceptance criteria (each - [ ] item becomes a task); labels (inform approach); related issues.

Impact Analysis

Grep for issue keywords and check git log --oneline -10 -- <related path>. Map acceptance criteria to likely file changes. Flag when the issue touches multiple modules (coordination), shared utilities (side-effect risk), or test fixtures (cross-suite updates).

Task Decomposition

For each acceptance criterion:

TaskCreate(
  subject: "Implement: {criterion summary}",
  description: "Acceptance criterion: {full text}\nLikely files: {paths}\nVerification: {how to check}"
)

Rules: at least one task per criterion; add infrastructure tasks (migrations, config) when needed; add a verification task at the end; set dependencies with addBlockedBy for sequential work.

Parallel Task Detection

Different files with no shared imports, or different directories → parallelizable. Same file → sequential dependency. If agent teams are enabled and >5 criteria have independent file sets → suggest team dispatch.

Decision Journal Init

Create {JOURNAL_DIR}/issue-{N}.md (default .decisions/):

# Decision Journal: Issue #{N} — {title}
**Issue**: #{N} | **Branch**: {branch-name} | **Started**: {YYYY-MM-DD}
---

The autonomous-workflow skill governs ongoing entry format (Init/Log/Summarize).

Verification

Branch setup is valid when: the branch name matches a convention pattern; issue details loaded; at least one task per acceptance criterion; no circular dependencies; journal initialized.

---name: branch-and-task-managementdescription: Create feature branches with naming conventions, load full issue context and impact analysis, and decompose acceptance criteria into atomic parallel tasks with dependencies. Use when starting work on a GitHub issue. This skill MUST be consulted because starting code without context causes misaligned implementations and wasted effort.---# Branch and Task Management## ContractIron law: no code before context — read the issue body and comments, load history, understand scope, then create the branch. Invoked by `/flow:start` (Phase 1 EXPLORE through Phase 2 PLAN) and by the `implementation-planner` agent for task decomposition. Returns: a feature branch named per `conventions.branchPatterns` checked out from the default branch, one TaskCreate task per acceptance criterion with dependencies set, and an initialized decision journal at `{JOURNAL_DIR}/issue-{N}.md`. Permitted skips: none — an issue without acceptance criteria is not decomposed; it goes back to the user (or `Skill(issue-crafting)`) first.## Pre-ConditionsBefore creating a branch, confirm: the issue exists and is OPEN; it has acceptance criteria (otherwise ask the user or create them); no existing branch already addresses it; you have read the full body AND comments.## Branch Creation```bashDEFAULT_BRANCH=$(gh repo view --json defaultBranchRef --jq '.defaultBranchRef.name' 2>/dev/null || echo "main")git fetch origin "$DEFAULT_BRANCH"git checkout -b "feature/issue-{N}-{desc}" "origin/$DEFAULT_BRANCH"```Patterns from `settings.json` → `conventions.branchPatterns`: `feature/issue-{N}-{desc}`, `fix/issue-{N}-{desc}`, `docs/issue-{N}-{desc}`. `{desc}` is 3-5 kebab-case words.## Issue Context LoadingFetch in parallel: `gh issue view $N --json title,body,labels,assignees,milestone` and `gh issue view $N --comments`. Extract: title; acceptance criteria (each `- [ ]` item becomes a task); labels (inform approach); related issues.## Impact AnalysisGrep for issue keywords and check `git log --oneline -10 -- <related path>`. Map acceptance criteria to likely file changes. Flag when the issue touches multiple modules (coordination), shared utilities (side-effect risk), or test fixtures (cross-suite updates).## Task DecompositionFor each acceptance criterion:```TaskCreate( subject: "Implement: {criterion summary}", description: "Acceptance criterion: {full text}\nLikely files: {paths}\nVerification: {how to check}")```Rules: at least one task per criterion; add infrastructure tasks (migrations, config) when needed; add a verification task at the end; set dependencies with addBlockedBy for sequential work.## Parallel Task DetectionDifferent files with no shared imports, or different directories → parallelizable. Same file → sequential dependency. If agent teams are enabled and >5 criteria have independent file sets → suggest team dispatch.## Decision Journal InitCreate `{JOURNAL_DIR}/issue-{N}.md` (default `.decisions/`):```markdown# Decision Journal: Issue #{N} — {title}**Issue**: #{N} | **Branch**: {branch-name} | **Started**: {YYYY-MM-DD}---```The autonomous-workflow skill governs ongoing entry format (Init/Log/Summarize).## VerificationBranch setup is valid when: the branch name matches a convention pattern; issue details loaded; at least one task per acceptance criterion; no circular dependencies; journal initialized.

Frequently asked questions