startWork
slash command
claude_code
Created by Prasham Trivedi
Copy This Prompt
---
allowed-tools: Bash(git:*), FileSystem, Bash(npm:*), Bash(yarn:*), Bash(build:*), Bash(test:*), Read, Write, Edit, MultiEdit, Bash(ls:*), Bash(git:*), Bash(tree:*),Bash(grep:*), Bash(cat:*), Bash(find:*), Bash(gh:*)
description: Start or continue working on a planned task with proper branch management
---
Start or continue working on a task.
## Arguments provided by user in $ARGUMENTS
- Directory Name: This is the name of the directory or task where
taskFindings.md file is located. This is used to understand the context of the
task and where to write the findings.
- Should work in same branch or separate branch.
## Autonomous Mode Detection
**IMPORTANT:** Operate in AUTONOMOUS MODE if `$ARGUMENTS` contains `mode: autonomous`.
In autonomous mode:
- Do NOT ask for user confirmation on branch strategy - use your recommendation
- Do NOT wait for user input - proceed immediately with best practices
- Do NOT ask user to run other commands - run them yourself or skip if not possible
- Make decisions based on complexity score and proceed
## Context and files
## Context
### Check if in git repository
- Is git repo: `test -d .git && echo "true" || echo "false"`
### Run these commands if you are in git repository
- Current git status:
`test -d .git && git status --porcelain || echo "Not a git repository"`
- Current branch:
`test -d .git && git branch --show-current || echo "Not a git repository"`
- Uncommitted changes:
`test -d .git && git diff --name-only || echo "Not a git repository"`
### This must always be available to you.
- Task findings: @taskNotes/{ArgumentDirectoryName}/taskFindings.md
## Steps
- Auto-detect if working directory is clean
- Smart branch naming with ticket prefixes
- Automatic dependency checking:
The **Dependency Auditor** skill will automatically check for outdated packages
and security vulnerabilities. Only warn if there are critical security issues
or major version changes, do not stop the flow.
- If taskFindings.md doesn't exist:
- In autonomous mode: STOP and report error - cannot proceed without plan
- In interactive mode: Ask user to run `/codePlanner` slash command
- Branch strategy decision (only if in git repository):
- For complexity 3 or less: working in same branch is fine
- For complexity 4 or more: working in separate branch is recommended
- Already in separate branch: continue in the same branch
**In autonomous mode:** Apply the recommendation automatically without asking.
**In interactive mode:** Ask the user to confirm.
- If separate branch is chosen/recommended and you are in git repository:
Make sure to check-out latest version of main, CREATE NEW branch with same
name as ${ArgumentDirectoryName}.
- If same branch is chosen or not in git repository:
Continue in the current branch. Ensure it is up-to-date with main if in git
repository.
- Before starting your work, write the current commit hash in
`taskNotes/{ArgumentDirectoryName}/currentCommitHash` file only if you are in
git repository.
- If the file already exists, that means the work is already started. Check the
content of `taskNotes/{ArgumentDirectoryName}` directory.
- Task Status from `taskNotesDirectory`
- `taskFindings.md`: Analysis done
- `currentCommitHash`: Work has started. Check commits between
currentCommitHash and latest commit.
- `backend-validation.md`: Backend work verified
- `frontend-validation.md`: Frontend work verified
- REMEMBER: Frontend don't start work untill verified backend contracts.
- Staged files and unstaged files can also show the progress.
- Complete the development of the task as per the findings Task Contents if
applicable.
- IMPORTANT: Use the **Conventional Commit** skill to generate commit messages.
Commit messages should not include any description - single line with emoji and
type only.
- Do not push the code yet, only keep the code locally (only applicable if in
git repository).
- Once the work is done, typechecked and tested:
**In autonomous mode:** Automatically run `/completeWork` to finalize.
**In interactive mode:** Ask the user to run `/completeWork` slash command.
### Checkpoint Strategy
- Commit after each MAJOR component (not minor changes)
- Checkpoints: BACKEND API, MCP Integration, FRONTEND Integration
- Micro checkpoints:
- Go up from Atomic unit of code to the complete Feature development, commit
after each major component.
- After Each Checkpoint:
- Use the **Checkpoint Validator** skill to:
- Run incremental typecheck
- Run tests only related to the changed files (not full test suite)
- Generate validation report
- At Final Checkpoint:
- Run typecheck before commits (not full build)
- Full test suite ONLY at final checkpoint
Slash Command Analysis ▶
- Requires arguments: Yes
- Agent references: None
- Skill references: None
- Analysis version: 1.0
Convert to Different Formats
Actions
(Clears cached conversions and forces re-processing)