$npx -y skills add krzysztofsurdy/code-virtuoso --skill ticket-deliveryEnd-to-end ticket workflow - from ticket analysis through investigation, planning, TDD implementation, committing, and PR creation. Adapts to your environment via a supplement file created on first run. Use when starting work on a ticket, implementing a feature end-to-end, or fol
| 1 | # Ticket Workflow |
| 2 | |
| 3 | End-to-end workflow that takes a ticket ID and drives it through analysis, codebase investigation, planning, TDD implementation, committing, and PR creation. Adapts to your stack via an interactive supplement created on first run. |
| 4 | |
| 5 | ## Phase 0: Environment Supplement |
| 6 | |
| 7 | Before anything else, check for a supplement file stored alongside this skill (e.g., `.supplement.md` in the skill directory). |
| 8 | |
| 9 | ### If supplement exists |
| 10 | Read the file and load all `$VARIABLE` values for use throughout the workflow. |
| 11 | |
| 12 | ### If supplement is missing |
| 13 | Run an interactive setup. Present an interactive prompt (or the platform's equivalent) for each question, wait for the answer, then proceed to the next. Save all answers to a supplement file stored alongside this skill (e.g., `.supplement.md` in the skill directory). |
| 14 | |
| 15 | **Questions to ask** (see [supplement reference](references/supplement-questions.md) for full details): |
| 16 | |
| 17 | 1. **Ticket system** - GitHub Issues / GitLab Issues / Jira / Linear / Other? Integration tool name for fetching tickets? |
| 18 | 2. **Version control** - Bitbucket / GitHub / GitLab? CLI tool (`gh`, `glab`)? Main branch name? |
| 19 | 3. **Error tracking** - Bugsnag / Datadog / Sentry / none? Integration tool name? |
| 20 | 4. **Logging** - CloudWatch / Datadog / ELK / Grafana / none? Tool access or manual paste? |
| 21 | 5. **Database** - ORM (ActiveRecord / Doctrine / Eloquent / Prisma / TypeORM / none)? Read-only query command? |
| 22 | 6. **CI/CD hooks** - Pre-commit hooks to skip? (e.g., `SKIP=lint,test git commit`) |
| 23 | 7. **Testing framework** - Go test / Jest / PHPUnit / pytest / RSpec / other? TDD style preference (London / Chicago)? |
| 24 | 8. **Architecture** - Monolith / microservices? Admin panel? API layer (REST / GraphQL / both)? |
| 25 | 9. **PR conventions** - Branch naming pattern? Commit message format? |
| 26 | |
| 27 | After saving, confirm: "Supplement saved. You can edit it anytime at the supplement file path or delete it to re-run setup." |
| 28 | |
| 29 | --- |
| 30 | |
| 31 | ## Delegation Strategy |
| 32 | |
| 33 | This workflow can use a multi-agent team when available. The coordinating agent delegates to specialized sub-agents for parallel investigation. |
| 34 | |
| 35 | ### Approach |
| 36 | | Strategy | Purpose | |
| 37 | |---|---| |
| 38 | | **Coordinating agent** | Orchestrates the full workflow, delegates investigation, reviews results | |
| 39 | | **Investigation sub-agents (x1-3)** | Explore specific codebase layers in parallel during Phase 3 | |
| 40 | | **Implementation sub-agent** | Executes TDD cycles during Phase 6 (or the coordinating agent does this directly) | |
| 41 | |
| 42 | Delegate to specialized sub-agents if available. Use delegation only when the ticket scope justifies parallel investigation. For small bug fixes, a single agent can handle everything alone. |
| 43 | |
| 44 | --- |
| 45 | |
| 46 | ## Phase 1: Ticket Intelligence Gathering |
| 47 | |
| 48 | ### Step 1.1 - Fetch Ticket Details |
| 49 | |
| 50 | **If `$TICKET_MCP_TOOL` is configured:** |
| 51 | ``` |
| 52 | Use $TICKET_MCP_TOOL to fetch ticket $TICKET_ID |
| 53 | ``` |
| 54 | |
| 55 | **If no integration tool is configured:** |
| 56 | Ask the developer: "Please paste the ticket title, description, and acceptance criteria." |
| 57 | |
| 58 | ### Step 1.2 - Parse Ticket Information |
| 59 | Extract and organize: |
| 60 | - **Title** and **description** |
| 61 | - **Acceptance criteria** (numbered list) |
| 62 | - **Ticket type** (bug / feature / improvement / refactor) |
| 63 | - **Priority** and **linked tickets** |
| 64 | - **Attachments or screenshots** (note them for later) |
| 65 | |
| 66 | ### Step 1.3 - Fetch Error Context (bugs only) |
| 67 | |
| 68 | **If `$ERROR_TRACKING_MCP_TOOL` is configured and ticket is a bug:** |
| 69 | ``` |
| 70 | Use $ERROR_TRACKING_MCP_TOOL to fetch error details linked in the ticket |
| 71 | ``` |
| 72 | Extract: stack trace, affected file paths, frequency, first/last seen. |
| 73 | |
| 74 | **If no integration tool is configured:** Ask developer to paste relevant error details or stack traces. |
| 75 | |
| 76 | ### Step 1.4 - Fetch Log Context (if applicable) |
| 77 | |
| 78 | **If `$LOGGING_TOOL` is configured:** |
| 79 | Fetch relevant log entries around the error timeframe. |
| 80 | |
| 81 | **If manual:** Ask developer to paste relevant log snippets if available. |
| 82 | |
| 83 | ### Step 1.5 - Summarize Findings |
| 84 | Present a structured summary: |
| 85 | ``` |
| 86 | TICKET: $TICKET_ID |
| 87 | TYPE: [bug|feature|improvement|refactor] |
| 88 | SUMMARY: [one-line summary] |
| 89 | ACCEPTANCE CRITERIA: [numbered list] |
| 90 | ERROR CONTEXT: [if bug - stack trace summary, affected files] |
| 91 | INITIAL HYPOTHESES: [2-3 possible approaches] |
| 92 | ``` |
| 93 | |
| 94 | Ask the developer: "Does this look correct? Any additional context before I investigate the codebase?" |
| 95 | |
| 96 | --- |
| 97 | |
| 98 | ## Phase 2: Branch Setup |
| 99 | |
| 100 | ```bash |
| 101 | git fetch origin |
| 102 | git checkout $MAIN_BRANCH |
| 103 | git pull origin $MAIN_BRANCH |
| 104 | ``` |
| 105 | |
| 106 | Do NOT create a feature branch yet - that happens in Phase 5 after planning. |
| 107 | |
| 108 | --- |
| 109 | |
| 110 | ## Phase 3: Codebase Investigation |
| 111 | |
| 112 | Investigate the codebase to understand the area affected by the tic |