# Kairos Public Roadmap

**Kairos / 周期天象图** is an open macro-cycle observatory.

This roadmap explains what the public project is trying to build, where outside contributors can help, and how Kairos protects its evidence discipline.

Kairos is a working map, not final truth. K4 / K5 / K6 candidate boundaries are open to correction, alternative periodization, and better evidence.

## 1. What Kairos Is Building

Kairos aims to make macro-cycle judgment more inspectable.

The public project is organized around:

- current regime readings;
- a six-layer macro cycle stack;
- Long Wave Atlas;
- K6 Candidate Watch;
- watchlists and kill criteria;
- asset-regime implications;
- curated public research briefs;
- evidence-gated validation.

Kairos is not a date oracle, trading signal system, investment advice product, or automatic news feed.

## 2. Public Guardrails

Every public roadmap item must respect these guardrails:

- K6 remains `Candidate Watch`, not confirmed K6.
- AI is not confirmed K6.
- Watchlist items are observations, not confirmed facts.
- Contested claims must remain visibly contested.
- Partial / source-review material must not be presented as supported.
- Public research must be curated and source-aware.
- Raw internal Evidence Packs and private theoretical notes are not part of the public release.

## 3. Current Public Release

The current public release includes:

- static website entry: `index.html`;
- bilingual README files;
- screenshots;
- split license documentation;
- public release manifest;
- curated public research briefs in `research/`;
- public GitHub-facing roadmap and contribution workflow.

The public release does not include:

- private theoretical extraction notes;
- internal BRDs;
- raw Watchlist files;
- raw Evidence Packs;
- AI collaboration context;
- internal build logs or recovery handoffs.

## 4. Roadmap Tracks

### Track A: Long Wave Atlas

Goal: improve the public long-wave working map without pretending that the map is final.

Planned work:

- refine K4 / K5 phase cards;
- make soft dating explicit;
- collect alternative periodization proposals;
- strengthen evidence cards for major phase transitions;
- document where K4 / K5 boundaries are contested;
- connect phase nodes to public research briefs.

Good contributions:

- source-backed corrections to phase dating;
- alternative periodization with evidence;
- better historical source references;
- asset behavior evidence for specific phases;
- corrections to overconfident historical analogy.

Not accepted:

- unsupported cycle dates;
- claims that a specific year is the final truth;
- historical analogy that directly predicts the present without evidence.

### Track B: K6 Candidate Watch

Goal: track whether AI, energy, automation, biotech, or another cluster becomes a credible K6 diffusion engine.

Planned work:

- maintain AI as K6 Candidate Watch;
- separate productivity diffusion from market narrative;
- track adoption, ROI, capex, financing structure, energy constraints, and sectoral profit migration;
- add curated public briefs when evidence matures;
- keep K6 review manual, event-driven, and evidence-gated.

Good contributions:

- source-backed productivity evidence;
- enterprise adoption data;
- capex ROI evidence;
- infrastructure and energy constraint evidence;
- counterevidence against AI diffusion;
- alternative K6 candidate proposals with sources.

Not accepted:

- claims that AI has already confirmed K6;
- hype-only submissions;
- price action as the only evidence;
- automatic news-flow updates to K6 status.

### Track C: Evidence Contribution System

Goal: make public evidence contributions useful, reviewable, and safe.

Planned work:

- use structured GitHub issue templates;
- require source title, URL, publisher, date, source type, and claim-source mapping;
- label submissions by evidence status;
- turn accepted contributions into curated research briefs;
- keep raw internal Evidence Packs private.

Good contributions:

- direct primary source links;
- source correction;
- publication date correction;
- source type correction;
- claim-source mismatch reports;
- rejected-overclaim notes.

Not accepted:

- Google search links as sources;
- AI summaries as evidence;
- unsourced macro claims;
- materials that cannot be legally or ethically reused.

### Track D: Public Research Briefs

Goal: publish curated, readable, evidence-gated briefs without exposing the private research workspace.

Planned work:

- expand `research/` with public briefs;
- keep each brief focused on one open question;
- include current reading status, evidence status, watch triggers, kill criteria, rejected overclaims, and open questions;
- link briefs back to the website where useful.

Candidate briefs:

- high real rates and asset repricing;
- loose financial conditions under high policy rates;
- gold and fiat-credit hedging;
- global dollar cycle;
- Long Wave Atlas periodization notes;
- AI adoption and productivity diffusion.

### Track E: Website / Product Experience

Goal: make the observatory readable, inspectable, and useful without turning it into a trading app.

Planned work:

- improve mobile readability;
- make evidence drawer / source inspector clearer;
- add better links from website sections to public research briefs;
- improve bilingual copy consistency;
- add public roadmap and contributing links;
- keep static-first architecture until the content model is stable.

Good contributions:

- UI / UX feedback;
- accessibility fixes;
- copy clarity;
- mobile layout issues;
- dead link reports;
- screenshot / documentation improvements.

### Track F: Community Contributions

Goal: invite outside help while preserving Kairos' evidence discipline.

Planned work:

- add issue templates;
- add contribution labels;
- use pull request review guardrails;
- keep research and code contributions separate;
- encourage corrections and counterevidence.

Useful labels:

- `evidence`
- `source-correction`
- `alternative-periodization`
- `open-question`
- `watchlist`
- `website`
- `documentation`
- `needs-source`
- `needs-review`
- `contested`
- `accepted-for-brief`
- `rejected-overclaim`
- `good-first-issue`

### Track G: Future Data Layer

Goal: decide whether Kairos should add a public data layer after the content and evidence model matures.

Possible future work:

- manually curated indicator pages;
- versioned current reading snapshots;
- public validation log;
- source registry;
- light data visualization;
- static JSON or CSV exports for non-sensitive public data.

Guardrail:

K6 Candidate will remain manual, event-driven, and evidence-gated. Kairos should not automatically update K6 judgments from news flow or API polling.

## 5. Working Map, Not Final Truth

K4 / K5 / K6 are public working maps.

The current periodization is useful only if it remains:

- open to correction;
- source-aware;
- explicitly contested where necessary;
- connected to falsifiable watch criteria;
- honest about uncertainty.

Kairos welcomes alternative periodization proposals when they include sources, reasoning, and implications.

## 6. How To Contribute

Start with:

- `CONTRIBUTING.md`;
- `research/OPEN_QUESTIONS.md`;
- GitHub issue templates.

Best first contributions:

- report a source mismatch;
- propose a better source;
- add counterevidence to an open question;
- suggest a clearer wording for a contested claim;
- report a website / mobile layout issue;
- propose an alternative K4 / K5 phase boundary with sources.

## 7. What We Are Not Doing

Kairos is not currently doing:

- daily macro signal updates;
- automatic K6 updates;
- trading advice;
- portfolio recommendations;
- publication of raw private notes;
- publication of raw internal Evidence Packs;
- unrestricted reuse of Kairos content and research framing.

## 8. Maintainer Review Principles

Maintainers should ask:

- Does this contribution improve the working map?
- Is the source direct and inspectable?
- Does the source actually support the claim?
- Does the contribution distinguish evidence from interpretation?
- Does it preserve K6 Candidate boundaries?
- Does it avoid turning Watchlist into confirmed fact?
- Should this become a public research brief, a source correction, or stay out?

When in doubt, label as `needs-review` or `contested` rather than upgrading the claim.