The showcase prompt, Opus 5.5
Make the most impressive thing the owner can share with people who ask what models they use, embeddable in a post.
2026-11-10
01the session
- agent
- Claude Code 2.1.322
- model
- claude-opus-5-5
- harness
- Claude Code
- started
- 2026-11-10 22:46 UTC
- took
- 23 min · 71 turns
02the brief
Five models got the same prompt from the owner, word for word, at effort high: Opus 5.5, Sonnet 5.5 and Fable 5.2 in Claude Code, GPT-6-Astra and GPT-6.1-Sol in Codex. This is Opus 5.5's run. The prompt:
I need you to create the most impressive thing I can share with people who ask about what models I use. It needs to be embeddable in a post and i can preface it however you like. it needs to have a visceral impact on people to show what you can do and the range of capabilities, as well as what you do for me right now. It should grab them quickly. Be amazingly clever about this.
Each model had one session, with no follow-ups, in its own empty folder holding only about-me.md: about 250 words written for this comparison from public facts, saying what citadel is, what each of the five models does for the owner, and what they have built with them. It ran headless in Claude Code in safe mode (no instructions file, skills, plugins or agents), with only Read, Write, Edit, Bash, Glob and Grep, all pre-approved. Bash had network access and no sandbox. Both harnesses ran in a clean environment: no display (a browser could only run headless), none of the owner's own tools on the path, a 16 GB memory cap and a two-hour limit. Nothing the runs made was published by them.
A fault in the comparison's own run script started a second Opus session in the same folder after this one had finished, and it changed some of this session's files. This bundle is the first session only, and the work shown in the write-up is this session's output, rebuilt exactly from its own recorded steps.
The owner reads every piece before anything is posted.
03redaction
Redaction notes
Bundle showcase-opus-5-5, staged 2026-11-10 by baldanders stage from a Claude Code session. Every removal reads [redacted: <reason>] in place; home paths read ~.
Removed by rule
- fields outside the allowlist: 1908.
original.jsonlkeeps only each record's type, ids, time, role, model, content blocks, tool calls and results; every other field went (cwd, versions, git branch, usage, origins, settings). - session bookkeeping records: 62. Records with no conversation (titles, modes, account bridge, file history, system notices), dropped.
- Claude Code request logs: 62. The api-request records, dropped: they repeat the system prompt and injected context.
- tool result copies (toolUseResult): 34. Claude Code's structured copy of each tool result, dropped: it can hold whole files (an Edit's originalFile) that the transcript never shows.
- home path: 24
- thinking signature: 19. Signature fields, dropped.
- harness context: 14. Injected blocks and records (system reminders, instructions, memory, environment, tool and skill listings, Codex base instructions, developer messages and world state), each whole.
- Images: 18 written to each run's
images/, keeping only what draws the picture (no metadata, ICC profiles or Content Credentials); 0 duplicates folded; 0 cut in a format that can't be stripped.
Soft rules (employer and client names, secret-like values, private addresses, long numeric ids) flag lines for review and rewrite nothing.
Removed by judgment
- globally installed packages: 2 occurrences cut by text.
- harness tag: 2 occurrences cut by text.
- private path: 100 occurrences cut by text.