Background research behind MISSION.md and vision.md. Two rounds, each run through two deep-research tools, plus hand verification. New to Python packaging? Start with the primer.
| File | What it is | Trust |
|---|---|---|
| round-1-landscape-compass.md | Snapshot of Python packaging today, comparison with other ecosystems, ranked gaps | High. Primary reference for round 1 |
| round-1-landscape-gemini.md | Same brief, different tool | Low. Agrees on headlines; recommendations and several facts are wrong (see notes) |
| round-2-evolution-compass.md | How tools replaced each other in JS and Python; what other ecosystems do best | High. Primary reference for round 2 |
| round-2-evolution-gemini.md | Same brief, different tool | Low. Largely restates round-1-compass, which it cites as source [4] |
| round-1b-landscape-compass.md | A second, independent run of the round 1 brief (Oct 2026). Intended as round 3, but the round 1 brief was used | High |
| round-1b-landscape-gemini.md | Same, from Gemini | Low (see notes) |
| round-1b-synthesis.md | What round 1b confirmed, what's new, where it disagrees with us | — |
| round-2b-evolution-compass-unanchored.md | The round 2 brief re-run without round 1 attached, to check for anchoring | High |
| round-3-agents-compass.md | Round 3: is an agent-specific angle authentic? Distribution formats, sandbox environments, evidence of pain | High |
| round-3-agents-gemini.md | Same brief, from Gemini. Truncated after the verdict | Low (incomplete) |
| round-3-and-2b-synthesis.md | What round 3 and 2b mean for bundleup | — |
| round-4-use-cases-compass.md | Round 4: who needs self-contained Python, what each platform requires, which output formats to offer | High |
| round-4-use-cases-gemini.md | Same brief, from Gemini | Low (stale platform facts, over-rated fits) |
| round-4-synthesis.md | Use-case lists, platform needs, output strategy | — |
| round-5-cli-compass.md | Round 5: CLI and library API style guide (attachments didn't arrive; bundleup specifics assumed) | High |
| round-5-cli-gemini.md | Same brief, from Gemini | Medium (agrees on rules; startup numbers not credible) |
| round-5-synthesis.md | Points to the resulting CLI style guide | — |
| round-6-next-tool-compass.md | Round 6: the next tool for shipping Python, through the JS ecosystem's history; gaps and complaints (run without attachments) | High on structure and primary sources; reaction counts unavailable, Reddit/X not searched |
| round-6b-next-tool-compass.md | Same brief, a second Compass run | Same; agrees with round 6 |
| round-7-agents-compass.md | Round 7: Python when agents write, run and consume the code (the sceptical run) | Medium-high; many 2026 arXiv preprints, unchecked |
| round-7b-agents-compass.md | Same brief, a second Compass run | Same; agrees with round 7 |
| round-6-and-7-synthesis.md | What rounds 6 and 7 mean for bundleup | — |
| verification-notes.md | Claims checked by hand, and known errors in each report | — |
| prompts.md | Index of the research briefs, one file per round in prompts/ | — |