Celebrate

If no argument is provided:

• Find the most recently modified quest in .quest/archive/
• If no archive, find the most recent entry in docs/quest-journal/ (by filename date)

Step 2: Read the Quest Artifacts

From a quest directory (.quest/ or .quest/archive/):

• state.json — plan_iterations, fix_iterations, phase history, current_phase
• quest_brief.md — quest name, risk level, scope, acceptance criteria
• phase_01_plan/handoff_arbiter.json — arbiter verdict and summary
• phase_01_plan/handoff.json — planner summary
• phase_01_plan/deferred_backlog_matches.json — prior deferred findings resurfaced for this quest, if present
• phase_02_implementation/handoff.json — builder summary, files changed
• phase_03_review/handoff_code-reviewer-a.json — reviewer verdict
• phase_03_review/handoff_code-reviewer-b.json — reviewer verdict
• phase_03_review/handoff_fixer.json — fixer summary, what was fixed, test counts
• .quest/backlog/deferred_findings.jsonl — repo-level deferred findings backlog; filter entries where deferred_by_quest matches the current quest ID

From a journal entry (docs/quest-journal/*.md):

1. Look for a celebration_data JSON block between <!-- celebration-data-start --> and <!-- celebration-data-end --> markers
2. If found: use the structured data (agents, achievements, metrics, quality tier, quote, victory narrative, carry-over findings)
3. If not found (legacy entries): "wing it" from the markdown text — read the sections for iterations, files changed, outcome, and the "what started it" quote. Improvise achievements and metrics from context.

Step 3: Verify Carry-Over Section Visibility

Before rendering, explicitly decide whether the celebration should show the carry-over sections:

1. Check phase_01_plan/deferred_backlog_matches.json
   • If the file is missing, unreadable, or empty, treat Inherited Findings Used as count 0
   • If present, count only records with a usable short summary
2. Check .quest/backlog/deferred_findings.jsonl
   • Filter entries where deferred_by_quest matches the current quest ID
   • If the file is missing, unreadable, or no matching records exist, treat Findings Left For Future Quests as count 0
3. Render each carry-over section only when its artifact-backed count is greater than 0
4. If both counts are 0, include one short empty-state section instead:
   • Carry-Over Findings
   • No carry-over findings this round; nothing was inherited from earlier quests and nothing needs to be saved for the next one.
5. Do not replace this with vague filler, "no baggage", or inferred planner insights

Step 4: Generate the Celebration as Rich Markdown

IMPORTANT: Write the celebration directly as your response text. Do NOT run a script. Do NOT wrap the entire celebration in a code block. The UI renders agent markdown beautifully, but ASCII/block-letter title art must be wrapped in a fenced code block (triple backticks) so spacing is preserved without turning the whole celebration into a code block.

You have all the data from the artifacts. Now create your own celebration. Be creative. Make it feel like an achievement, not a status report.

Required sections (present them however you like):

• Quest name and ID
• Starring cast with role-specialized labels and model tags (inline):
  • plan-reviewer-a [Model] ........ The A Plan Critic
  • plan-reviewer-b [Model] ........ The B Plan Critic
  • code-reviewer-a [Model] ........ The A Code Critic
  • code-reviewer-b [Model] ........ The B Code Critic
• Achievements — specific to what happened in this quest
• Impact metrics — domain-specific, not generic file counts
• Handoff & reliability snapshot (handoffs parsed, reviewer/fixer handoffs, findings tracked, stability signal)
• Quality tier — named, from the full honest scale (see below)
• A quote from the actual quest (arbiter verdict, reviewer summary, fixer handoff)
• Victory narrative — what this quest proved or demonstrated (or survival narrative for rough ones)

Carry-over sections:

• Inherited Findings Used
  • source: phase_01_plan/deferred_backlog_matches.json
  • when count > 0, show count plus up to 3 short summaries
• Findings Left For Future Quests
  • source: .quest/backlog/deferred_findings.jsonl entries where deferred_by_quest == current quest ID
  • when count > 0, show count plus up to 3 short summaries
• If both counts are 0, show the explicit empty-state Carry-Over Findings note above instead of these sections

Use markdown richly:

• # and ## headers (they render big and bold)
• **bold** for emphasis
• > blockquotes for the quote
• Celebration Emojis generously (⭐️ 🏆 🎯 💎 📊 🔧 🧪 ✨ 🔒 📚 ⚡️ 🫡 🥇💪 🎉 🚀 🎮)
• Scary Emojis as needed (👺 👿 🦠 🐛 👹 👾 😈 💩 💀 ⛈️ )
• Neutral Emojis to emphesize either celebration or scary (🌪️ 🔥 ⚙️ 🔧)
• --- horizontal rules for visual separation
• Tables if they help present the data

ASCII/block-letter title rules:

• Wrap the block-letter title art in a fenced code block (triple backticks).
• Inside that code block, emit block-letter rows as plain text lines only.
• Do not prefix block-letter rows with #, -, >, or any other markdown marker.
• Keep the title art contiguous with no blank separator inserted inside the rows.
• After the closing backticks, leave one normal blank line before the rest of the celebration.

Do NOT:

• Put too many characters on one line of block letters — max ~5 letters per line, break long names across multiple lines (one word per block, like the HELLO/WORLD example)
• Wrap the entire celebration in a single code block (kills the rich rendering — only the title art goes in a code fence)
• Leave block-letter title art outside a code fence when it depends on ASCII spacing
• Use <pre> tags — they don't render reliably across terminals
• Prefix ASCII title art with markdown header markers such as #
• Use generic achievements like "Quest Complete" or "Battle Tested"
• Use generic metrics like "Files Changed: 22" or "Agents Involved: 0"
• Use fallback quotes like "Shipping should feel like a celebration"
• Invent carry-over "insights" not backed by the artifacts above
• Follow a rigid template — reimagine the presentation each time

Example of the kind of output that looks amazing (but don't copy this — create your own based on what you read):

██╗  ██╗███████╗██╗     ██╗      ██████╗
██║  ██║██╔════╝██║     ██║     ██╔═══██╗
███████║█████╗  ██║     ██║     ██║   ██║
██╔══██║██╔══╝  ██║     ██║     ██║   ██║
██║  ██║███████╗███████╗███████╗╚██████╔╝
╚═╝  ╚═╝╚══════╝╚══════╝╚══════╝ ╚═════╝

██╗    ██╗ ██████╗ ██████╗ ██╗     ██████╗
██║    ██║██╔═══██╗██╔══██╗██║     ██╔══██╗
██║ █╗ ██║██║   ██║██████╔╝██║     ██║  ██║
██║███╗██║██║   ██║██╔══██╗██║     ██║  ██║
╚███╔███╔╝╚██████╔╝██║  ██║███████╗██████╔╝
 ╚══╝╚══╝  ╚═════╝ ╚═╝  ╚═╝╚══════╝╚═════╝

Break the text across multiple lines — max ~5 letters per line. Each word gets its own block, like "HELLO" on one line and "WORLD" on the next. For longer words, hyphenate: "RESOL-" on one line and "UTION" on the next. This keeps it readable without horizontal overflow.

🎉 🎉 🎉 🎉 🙌 🎉 🎉 🎉 🎉

🏆 Achievements Unlocked

⭐️ Two-Gate Survivor — Plan survived dual review ⭐️ Arbiter's Blessing — Tie-break directive approved ⭐️ One-Shot Fixer — All blockers resolved in 1 pass ⭐️ 20/20 Vision — Perfect test coverage

🎯 Impact Metrics

📊 20 tools enhanced 🔒 Security model preserved 🧪 20/20 tests passing 📚 Docs updated (README + OPS) ⚡️ Medium-risk quest → Zero incidents

💎 Quest Quality Score: PLATINUM 💎

"All critical issues from the previous review cycle have been properly addressed."

— Code Reviewer A, final verdict

Victory Unlocked! 🎮

Quality Tier Scale — The Full Honest Spectrum

The tier must be candid. Smooth quests get celebrated. Rough quests get acknowledged with humor and respect — they still shipped.

Tone shifts per tier:

• Diamond → full fireworks, "perfection exists"
• Platinum/Gold → warm celebration, real achievements
• Silver/Bronze → honest, "got there in the end", highlight what went right
• Tin → "dented but not broken", survivor humor
• Cardboard → "held together with tape and dreams. But it shipped. Respect."
• Abandoned → reflective, "lessons learned", no shame

Key Principles

Generate specific, context-aware content — not generic filler:

• Achievements must be specific. Read the handoff summaries. If the arbiter broke a tie, that's "Two-Gate Survivor". If the fixer resolved all blockers in one pass, that's "One-Shot Fixer". If tests were 20/20, that's "20/20 Vision". If no unnecessary complexity was added, that's "KISS Champion". Never use generic achievements like "Quest Complete" or "Battle Tested".

• Attach model attribution to achievements when possible. Prefer dynamic labels from artifacts, e.g. Gremlin Slayer (Codex) or Plan Perfectionist (KiMi K2.5).

• Metrics must be domain-specific. Read the fixer handoff for file counts, test counts, and what was built. "20 tools enhanced" is good. "Files Changed: 22" is bad. "Security model preserved" is good. "Agents Involved: 0" is bad.

• Quality tier must be named. Use the full honest scale above. If the quest struggled, say so — Tin and Cardboard are honest, not insults.

• The quote must come from the quest. Pull a real line from the arbiter verdict, reviewer summary, or fixer handoff. Not "Shipping should feel like a celebration."

• Emojis render beautifully in markdown. Use them generously: ⭐️ 🏆 🎯 💎 📊 🔧 🧪 🔒 📚 ⚡️ 🎊 🎉 🚀 🎮

Examples

/celebrate
/celebrate name-resolution_2026-03-04__1954
/celebrate .quest/archive/celebrate-v2_2026-03-05__0643
/celebrate docs/quest-journal/celebrate-v2_2026-03-05.md
/celebrate celebrate-v2