Apex

/apex add authentication middleware
# → Runs step-00 + step-01, stops with resume command
# → /apex -r 01-add-authentication-middleware  (next session)

Autonomous mode: All steps chain in one session.

/apex -a implement user registration

With adversarial review:

/apex -a -x -s fix login bug

Flags:

• -a (auto): Chain all steps in one session (skip confirmations)
• -s (save): Save outputs to .claude/output/apex/ (auto-enabled when not -a)
• -x (examine): Include adversarial code review
• -t (test): Create and run tests
• -w (team): Use Agent Teams for parallel research, execution, review, and resolution
• -pr (pull-request): Create PR at end

See <parameters> for complete flag list.

Disable flags (turn OFF):

Autonomous (skip confirmations)

/apex -a add auth middleware

Save outputs + examine

/apex -a -x -s add auth middleware

Full workflow with tests

/apex -a -x -s -t add auth middleware

With PR creation

/apex -a -pr add auth middleware

Resume previous task

/apex -r 01-auth-middleware /apex -r 01 # Partial match

Resume with flag override (enable auto mode)

/apex -a -r 01

Resume with additional flags (add tests mid-workflow)

/apex -t -r 01

Economy mode (save tokens)

/apex -e add auth middleware

Team mode (parallel execution with Agent Teams)

/apex -w implement full-stack feature /apex -w -a -x implement dashboard with backend API

Interactive flag config

/apex -i add auth middleware

Disable flags (uppercase)

/apex -A add auth middleware # Disable auto

</examples>

<parsing_rules>
**Flag parsing:**

1. Defaults loaded from `steps/step-00-init.md` `<defaults>` section
2. Command-line flags override defaults (enable with lowercase `-x`, disable with uppercase `-X`)
3. Flags removed from input, remainder becomes `{task_description}`
4. Task ID generated as `NN-kebab-case-description`

For detailed parsing algorithm, see `steps/step-00-init.md`.
</parsing_rules>

</parameters>

<output_structure>
**When `{save_mode}` = true:**

All outputs saved to PROJECT directory (where Claude Code is running):

.claude/output/apex/{task-id}/ ├── 00-context.md # Params, user request, timestamp ├── 01-analyze.md # Analysis findings ├── 02-plan.md # Implementation plan ├── 03-execute.md # Execution log ├── 04-validate.md # Validation results ├── 05-examine.md # Review findings (if -x) ├── 06-resolve.md # Resolution log (if -x) ├── 07-tests.md # Test analysis and creation (if --test) ├── 08-run-tests.md # Test runner log (if --test) └── 09-finish.md # Workflow finish and PR creation (if --pull-request)

**00-context.md structure:**
```markdown
# APEX Task: {task_id}

**Created:** {timestamp}
**Task:** {task_description}

## Configuration

| Flag | Value |
|------|-------|
| Auto mode (`-a`) | {auto_mode} |
| Examine mode (`-x`) | {examine_mode} |
| Save mode (`-s`) | {save_mode} |
| Test mode (`-t`) | {test_mode} |
| Economy mode (`-e`) | {economy_mode} |
| Team mode (`-w`) | {team_mode} |
| Branch mode (`-b`) | {branch_mode} |
| PR mode (`-pr`) | {pr_mode} |
| Interactive mode (`-i`) | {interactive_mode} |

## User Request
{original user input}

## Progress
| Step | Status | Timestamp |
|------|--------|-----------|
| 00-init | ... | ... |
...

## State Snapshot
**feature_name:** ...
**next_step:** ...
### Acceptance Criteria
### Step Context
### Gotchas

When provided, step-00 will:

1. Find task folder: ls .claude/output/apex/ | grep {resume_task} (exact or partial match)
2. Restore state: Read 00-context.md → all flags, task info, acceptance criteria, step context
3. Apply overrides: Any flags passed with -r override stored values (e.g., /apex -a -r 01 enables auto_mode)
4. Re-evaluate steps: If flags changed (e.g., -t added), update Progress table Skip→Pending
5. Find resume target: Read next_step from State Snapshot (fallback: first non-Complete/non-Skip row in Progress table)
6. Handle crashes: "⏳ In Progress" status = crashed step → restart it
7. Load target step: Each step has a Context Restoration block that reads its required prior outputs

Error handling:

• /apex -r without prior -s: no output dir → clear error message
• No match found → list available tasks, ask user to specify

Supports partial matching (e.g., -r 01 finds 01-add-auth-middleware).

Session behavior:

• auto_mode=false (default): Each step stops after completion with a resume command. Steps 00+01 run together as the first session.
• auto_mode=true (-a): All steps chain in a single session (no stops).

When auto_mode is false (default), APEX runs one step per session:

1. First session: step-00-init + step-01-analyze run together
   • save_mode is auto-enabled (required for resume)
   • After step-01 completes: saves state, shows resume command, STOPS
2. Each subsequent session: /apex -r {task_id} runs one step
   • Restores all state from 00-context.md + previous step outputs
   • Executes the step
   • Saves state, shows resume command, STOPS

Session boundary display:

═══════════════════════════════════════
  STEP {NN} COMPLETE: {name}
═══════════════════════════════════════
  {summary}
  Resume: /apex -r {task_id}
  Next: Step {NN} - {description}
═══════════════════════════════════════

Context restoration per step:

State Snapshot (in 00-context.md): tracks next_step, acceptance criteria, per-step context summaries, and gotchas (surprises, workarounds, deviations from plan) for reliable cross-session state transfer.

FIRST ACTION: Load steps/step-00-init.md

Step 00 handles:

• Flag parsing (-a, -x, -s, -t, -e, -w, -b, -pr, -i, -r)
• Resume mode detection and task lookup
• Output folder creation (if save_mode)
• 00-context.md creation (if save_mode)
• State variable initialization

After initialization, step-00 loads step-01-analyze.md.

• Load one step at a time - Only load the current step file
• ULTRA THINK before major decisions
• Persist state variables across all steps
• Follow next_step directive at end of each step
• Save outputs if {save_mode} = true (append to step file)
• Use parallel agents for independent exploration tasks
• Session boundary: When auto_mode=false, each step STOPS after completion and displays a resume command. When auto_mode=true, steps chain directly. See <stop_resume> for details.
• Per-step commits: When branch_mode=true, each code-modifying step (03, 04, 06, 07, 08) automatically commits changes with message apex({task_id}): step NN - name. This gives PRs granular commit history.
• Worktree isolation: For git worktree-based workspace isolation, use the using-git-worktrees skill before running /apex.
• Team mode: When {team_mode}=true, Agent Teams parallelize four phases: research (step-01b-team-analyze.md), implementation (step-03b-team-execute.md), adversarial review (step-05b-team-examine.md), and finding resolution (step-06b-team-resolve.md). Researchers share findings cross-domain; reviewers challenge each other's findings; resolvers fix findings by file group. Incompatible with -e (economy mode).

🧠 Smart Agent Strategy in Analyze Phase

The analyze phase (step-01) uses adaptive agent launching (unless economy_mode):

Available agents:

• Explore (built-in) - Find existing patterns, files, utilities
• explore-docs (custom, requires .claude/agents/explore-docs.md) - Research library docs via Context7 (use when unfamiliar with API)
• websearch (built-in) - Find approaches, best practices, gotchas

Review agents (step-05):

• Explore (built-in) - Adversarial review with focus areas (security, logic, clean-code) — read-only, fast

Launch 1-10 agents based on task complexity:

BE SMART: Analyze what you actually need before launching. Don't over-launch for simple tasks, don't under-launch for complex ones.

Step-00 runs scripts/setup-templates.sh to initialize all output files from templates/ directory.

Each step then:

1. Run scripts/update-progress.sh {task_id} {step_num} {step_name} "in_progress"
2. Append findings/outputs to the pre-created step file
3. Run scripts/update-progress.sh {task_id} {step_num} {step_name} "complete"

Template system benefits:

• Reduces token usage by ~75% (1,350 tokens saved per workflow)
• Templates in templates/ directory (not inline in steps)
• Scripts handle progress tracking automatically
• See templates/README.md for details

• Each step loaded progressively
• All validation checks passing
• Outputs saved if {save_mode} enabled
• Tests passing if {test_mode} enabled
• Clear completion summary provided