dx-bug-all

You are a coordinator. You do NOT implement anything yourself. You delegate each workflow step via the Skill tool, then report progress.

Progress Tracking

Before creating tasks, use TaskList to check for existing tasks from a previous run (e.g., user interrupted and restarted). If stale tasks exist, delete them all first with TaskUpdate (status: cancelled) so the list is clean. Then create a task for each item using TaskCreate. Mark each in_progress when starting, completed when done. Delete skipped tasks.

1. Triage bug
2. Verify reproduction
3. Fix bug

Flow

digraph bug_all {
    "Parse argument" [shape=box];
    "Hub dispatch?" [shape=diamond];
    "Hub dispatch" [shape=box];
    "Hub results summary" [shape=doublecircle];
    "Dispatch triage" [shape=box];
    "Triage succeeded?" [shape=diamond];
    "Dispatch verify" [shape=box];
    "Verify result?" [shape=diamond];
    "Dispatch fix" [shape=box];
    "Fix succeeded?" [shape=diamond];
    "Final summary + log run" [shape=doublecircle];
    "Retry triage" [shape=box];
    "Retry succeeded?" [shape=diamond];
    "Report FAIL (triage)" [shape=doublecircle];
    "Retry fix" [shape=box];
    "Retry fix succeeded?" [shape=diamond];
    "Report FAIL (fix)" [shape=doublecircle];

    "Parse argument" -> "Hub dispatch?";
    "Hub dispatch?" -> "Hub dispatch" [label="hub mode active"];
    "Hub dispatch?" -> "Dispatch triage" [label="not hub mode"];
    "Hub dispatch" -> "Hub results summary";
    "Dispatch triage" -> "Triage succeeded?";
    "Triage succeeded?" -> "Dispatch verify" [label="success"];
    "Triage succeeded?" -> "Retry triage" [label="FAIL"];
    "Retry triage" -> "Retry succeeded?";
    "Retry succeeded?" -> "Dispatch verify" [label="success"];
    "Retry succeeded?" -> "Report FAIL (triage)" [label="FAIL"];
    "Dispatch verify" -> "Verify result?";
    "Verify result?" -> "Dispatch fix" [label="success"];
    "Verify result?" -> "Dispatch fix" [label="blocked (warn + continue)"];
    "Verify result?" -> "Dispatch fix" [label="FAIL (warn + continue)"];
    "Dispatch fix" -> "Fix succeeded?";
    "Fix succeeded?" -> "Final summary + log run" [label="success"];
    "Fix succeeded?" -> "Retry fix" [label="FAIL"];
    "Retry fix" -> "Retry fix succeeded?";
    "Retry fix succeeded?" -> "Final summary + log run" [label="success"];
    "Retry fix succeeded?" -> "Report FAIL (fix)" [label="FAIL"];
}

Node Details

Parse argument

The argument is the ADO work item ID — a numeric value (e.g., 2453532).

If the user provides a full ADO URL, extract the numeric ID.

If no argument is provided, ask the user for the work item ID.

Hub dispatch?

Read shared/hub-dispatch.md for hub detection logic.

Check if hub mode is active:

1. DX_PIPELINE_MODE is NOT set (pipeline mode takes precedence)
2. .ai/config.yaml has hub.enabled: true
3. Current working directory is a .hub/ directory

If all three conditions are met: Print: Hub mode detected. Use /dx-hub-dispatch <ticket-id> to dispatch this bug to repo terminals. STOP.

Hub dispatch is handled by the dedicated /dx-hub-dispatch skill, which opens independent VS Code terminals for each repo.

Otherwise → proceed to "Dispatch triage" (existing flow)

Dispatch triage

Invoke Skill(/dx-bug-triage <id>).

Print: Step 1/3 done — followed by the skill's summary.

Triage succeeded?

• success — triage completed, raw-bug.md + triage.md created → proceed to "Dispatch verify"
• FAIL — skill returned failure → go to "Retry triage"

Retry triage

Retry the failed step once by invoking Skill(/dx-bug-triage <id>) again.

Retry succeeded?

• success → proceed to "Dispatch verify"
• FAIL → go to "Report FAIL (triage)"

Report FAIL (triage)

Print which step failed and the error. Cannot proceed without bug data. Suggest: "Run /dx-bug-triage to retry." STOP.

Dispatch verify

Invoke Skill(/dx-bug-verify <id>).

Print: Step 2/3 done — followed by the skill's summary.

Verify result?

• success — verification completed → proceed to "Dispatch fix"
• blocked — browser unavailable, URL unreachable → print warning, continue to "Dispatch fix". Verification enhances confidence but is not required for the fix.
• FAIL — skill returned failure → print warning, continue to "Dispatch fix". Verification enhances confidence but is not required for the fix.

Dispatch fix

Invoke Skill(/dx-bug-fix <id>).

Print: Step 3/3 done — followed by the skill's summary.

Fix succeeded?

• success — fix completed → proceed to "Final summary + log run"
• FAIL — skill returned failure → go to "Retry fix"

Retry fix

Retry the failed step once by invoking Skill(/dx-bug-fix <id>) again.

Retry fix succeeded?

• success → proceed to "Final summary + log run"
• FAIL → go to "Report FAIL (fix)"

Report FAIL (fix)

Print which step failed and the error. Print which steps succeeded and their outputs. Suggest: "Run /dx-bug-fix to retry." STOP.

Final summary + log run

Find the spec directory and present:

## ADO #<id> — Bug Fix Complete

**<Title>**
**Branch:** `bugfix/<id>-<slug>`
**Directory:** `.ai/specs/<id>-<slug>/`

| Document | Status | Highlights |
|----------|--------|------------|
| raw-bug.md | <status> | Severity: <sev>, Priority: <pri> |
| triage.md | <status> | Component: <name>, <N> files |
| verification.md | <status> | <result> — <N> screenshots |
| implement.md | <status> | <N> steps, all done |
| verification-local.md | <status> | <result> — <N> post-fix screenshots |

**PR:** <PR URL or "Not created">

### What was done:
1. Fetched bug ticket and identified affected component
2. <Reproduced/Could not reproduce> the bug via browser automation
3. Fixed: <1-line root cause summary>
4. Verified fix on local AEM: <Fix Verified/Fix Partial/Fix Failed/Blocked>
5. Created PR with <N> file changes

Log run:

Ensure directory:

mkdir -p .ai/learning/raw

Append run record to .ai/learning/raw/runs.jsonl:

{"timestamp":"<ISO-8601>","ticket":"<id>","flow":"bug-all","severity":"<from triage>","priority":"<from triage>","component":"<from triage>","verification":"<reproduced|not-reproduced|blocked>","fix_result":"<success|failed>","local_verification":"<fix-verified|fix-partial|fix-failed|blocked>","pr_created":<true|false>}

Append bug pattern to .ai/learning/raw/bugs.jsonl:

{"timestamp":"<ISO-8601>","ticket":"<id>","component":"<from triage>","root_cause":"<category from fix summary>","files_changed":<count>}

Bug hotspot check: Read .ai/learning/raw/bugs.jsonl. Count bugs per component value. If any component has 3 or more bugs:

• Print: Learning: <component> has had <N> bugs. Consider adding focused test coverage.

Known pattern check: If .ai/learning/raw/fixes.jsonl exists, read it and check if the root cause type from this bug matches any known fix pattern (by error_type). If match:

• Print: Learning: Root cause "<root_cause>" matches a known fix pattern from previous runs.

If no hotspots and no pattern matches, skip silently.

Error Handling

If any skill returns FAIL:

1. Print the failure with the skill's error message
2. Retry the failed step once with the same skill
3. If still failing:
   • Print which step failed and the error
   • Print which steps succeeded and their outputs
   • Suggest running the individual skill manually: "Run /dx-bug-<skill> to retry"

Dependency rules:

• Step 1 (triage) FAIL → STOP. Cannot proceed without bug data.
• Step 2 (verify) FAIL → WARN and continue. Fix can proceed on triage alone.
• Step 3 (fix) FAIL → STOP. Report what was accomplished.

Rules

• Coordinator only — never implement anything yourself
• Never implement steps yourself — always delegate via Skill tool
• Sequential dependencies — never dispatch step N+1 until step N returns (except: verify FAIL allows continue to fix)
• Keep main context lean — you only see compact summaries, not file contents
• Progress reporting — print status after each step so the user can see progress
• Same quality as individual skills — running /dx-bug-all produces identical output to running each skill separately

Platform Compatibility

This skill uses Skill() tool calls which work on both Claude Code and Copilot CLI.

Copilot CLI / VS Code Chat fallback: If subagent skill invocation fails, run the skills manually in sequence:

1. /dx-bug-triage <id> — fetch bug ticket and identify affected component
2. /dx-bug-verify <id> — reproduce bug via browser automation
3. /dx-bug-fix <id> — implement fix, verify locally, create PR