docs(strategy): Platform & Business Strategy 2026-07 — Polar.sh, EU AI Act, phased openness

[frankxai]

What

Adds docs/strategy/ARCANEA_PLATFORM_BUSINESS_STRATEGY_2026-07.md — a decision document covering:

• Polar.sh verdict: adopt as merchant of record for first-party digital revenue; keep the existing Stripe code (apps/web/app/api/stripe/* in arcanea-ai-app) dormant as the future Stripe Connect marketplace rail.
• Legal: EU AI Act Article 50 transparency obligations — C2PA content marking, AI-interaction disclosure, and AI-influencer persona disclosure plan. GDPR + DSA readiness notes.
• Business model simulation: four models scored; chosen stack = subscription + credits (Higgsfield-informed engine), OSS funnel (Vercel-informed), publishing rev-share as Phase 2–3 expansion.
• Phased openness: Sovereign (Frank solo) → Court of Creators (invite-only) → Open Academy (self-serve), with explicit gates between phases.
• Repo → product map, agent-swarm org chart, Q3/Q4 2026 OKRs, and a 30/60/90 plan.

Why

Grounded in the actual repo state: Stripe tiers + credit packs are already coded, arcanea-mcp / publishing-house-mcp / world-sdk packages exist, and the Feb 2026 two-domain strategy stands. This doc makes the monetization and openness decisions explicit and sets the compliance work against the Article 50 date.

Timing — read before merging

The doc was drafted 2026-07-13 and reads forward from that date. It is now landing after the Article 50 date of 2026-08-02, which has passed. A merge-time status note at the top of the doc states this plainly: the C2PA marking, AI-interaction disclosure, and AI-influencer labeling work described in §2 is a live obligation and unshipped, not a deadline to plan toward, and the counsel review tracked under OKR O2 is overdue. §8's 30/60/90 dates are measured from 2026-07-13, not from merge.

CI note

CLI Tests and Test Summary are red. They are not caused by this PR — it changes three files (.gitignore and two markdown docs, +190 lines) and contains no source code. The Arcanea Test Suite workflow has concluded failure on every run on main going back to 2026-03-01 (175 failures / 230 cancelled out of 1868 tests). Every other check is green: TypeScript, Security Audit, Production Build, TypeScript & Package Tests, No Split-Brain Publish, Quality Status, claude-review. That standing suite failure is worth its own triage, separately from this doc.

Not tracked elsewhere

The Q3/Q4 OKRs (O1–O4) and the overdue Article 50 items exist only in this document. If they should be falsifiable, they need a tracking issue — deliberately not created here, since that's a workflow call for the repo owner.

🤖 Generated with Claude Code

https://claude.ai/code/session_01EN5jWvgRUdkwUUnH3Ggcy9

---

Generated by Claude Code

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
arcanea-2	Error		Jul 13, 2026 1:30am
arcanea-web	Ready	Preview, Comment, Open in v0	Jul 13, 2026 1:30am

[coderabbitai]

Important

Review skipped

Draft detected.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: 2978c9d4-5997-403e-bd11-63fac6f7c347

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch claude/arcanea-strategy-platform-r5zyme

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[gemini-code-assist]

Code Review

This pull request introduces a comprehensive business and platform strategy document for Arcanea (July 2026), detailing decisions on monetization (adopting Polar.sh), legal compliance (EU AI Act), product architecture, and a phased rollout plan. The review feedback suggests minor editorial improvements, including correcting typos of 'FrankX' to 'frankxai' or 'Frank's', clarifying ambiguous phrasing regarding Web3 sequencing, and replacing the awkward noun 'irreversibles' with 'irreversible actions' to improve readability.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[gemini-code-assist]

The phrase "one more reason web3 sequences last" is slightly ambiguous and could be misread as "web3 sequences endure/last". To align with the terminology in Section 0 ("Sequence it last"), consider rephrasing this to "one more reason to sequence web3 last" or "one more reason web3 is sequenced last".

[gemini-code-assist]

The reference to "FrankX" appears to be a typo. Based on the repository strategy and GitHub organization name ("frankxai"), this should be updated to "frankxai" or "frankxai's" to maintain consistency.

[frankxai]

"FrankX" is intentional, not a typo — it's the brand name (frankx.ai) and the name of the sibling repo whose lib/gen layer is being referenced (same applies to the line-118 comment). The GitHub org handle is frankxai, but the brand and repo are "FrankX". The two phrasing suggestions (web3 sequencing, "irreversibles") were valid and are applied in 2d54be6.

---

Generated by Claude Code

[gemini-code-assist]

The reference to "FrankX" in "FrankX gen rules" appears to be a typo or inconsistent with the rest of the document (which uses "Frank" or "frankxai"). Consider updating it to "Frank's" or "frankxai's".

[gemini-code-assist]

The term "irreversibles" is slightly awkward as a noun here. Using "irreversible actions" or "irreversible operations" would be clearer and more professional.

Suggested change

	Structure the swarms as a cabinet with human-gated irreversibles (posting, sending, spending stay with Frank per protection-layer doctrine):
	Structure the swarms as a cabinet with human-gated irreversible actions (posting, sending, spending stay with Frank per protection-layer doctrine):

[claude]

Review: docs(strategy) — Platform & Business Strategy 2026-07

This is a single-file, docs-only PR (docs/strategy/ARCANEA_PLATFORM_BUSINESS_STRATEGY_2026-07.md, +167/-0). Since there's no source code, I reviewed it as a decision document rather than against the usual code checklist.

Overall
Well-structured and decisive: the TL;DR table up top, explicit what-NOT-to-do list, and phased gates (Sovereign -> Court -> Academy) with concrete exit criteria are good practice for a doc meant to drive decisions, not just describe options. The payments doctrine ('no autonomous money movement, ever — agents draft and verify; Frank approves capital') is a good safety guardrail to have written down explicitly.

Verifiability of repo-grounded claims
The doc says it is 'grounded in the actual state of the repos (arcanea, arcanea-ai-app, Starlight-Intelligence-System, ACOS).' I could only check the arcanea repo checked out here:

• Confirmed: packages/arcanea-mcp exists, matching the MCP claim.
• Confirmed: .arcanea/lore/CANON_LOCKED.md exists, matching the canon reference.
• Minor discrepancy: Section 5's product-surface table cites 'book/ (17 collections)', but book/ in this repo currently has 20 collection directories (plus README). Worth a quick correction, or a note that the count grew after the doc was drafted — small factual claims like this are the first thing readers spot-check.
• Unverifiable here: apps/web/app/api/stripe/*, CREDIT_PACKS, publishing-house, publishing-house-mcp, world-sdk, arcanea-registry-mcp, world-engine, and a design-system package don't exist in this repo. They likely live in arcanea-ai-app or another sibling repo, consistent with the doc's stated multi-repo scope, but it means the Polar/Stripe integration claims (e.g. 'Stripe routes already 503 gracefully without keys') aren't verifiable from this PR alone. Not a blocker, just worth flagging so 'grounded in actual repo state' isn't read as fully verified vs. true as of drafting time.

Compliance content
The Article 50 (Aug 2, 2026) deadline and the transparency/GDPR/DSA breakdown read as internally consistent and appropriately scoped (deployer/integrator of GPAI, not a provider). Since this is time-sensitive and legally consequential, a one-line note on who/what validated the Aug 2 date and the 'no high-risk category applies' conclusion (external counsel vs. internal read) would help whoever owns follow-through on the O2 OKR.

Nits
Consider a 'last verified' line near the top if this is meant to be a living doc that gets re-checked against arcanea-ai-app state over time, rather than a point-in-time snapshot.

Test coverage / security / performance
N/A — no code, no secrets introduced, nothing executable.

Verdict: Reasonable to merge as a decision record. The book-collection-count nit is easy to fix; the cross-repo claims are worth a mental asterisk rather than a blocking change.

[claude]

Review: PR #82 — Platform & Business Strategy 2026-07

This is a docs-only PR (single new file, docs/strategy/ARCANEA_PLATFORM_BUSINESS_STRATEGY_2026-07.md, +171/-0). Standard code-quality/perf/security/test-coverage criteria mostly don't apply here, so this review focuses on content accuracy, internal consistency, and repo hygiene.

What checks out

• EU AI Act timeline is accurate. Article 50 transparency obligations (Chapter IV) do apply from 2026-08-02 (24 months after entry into force), separate from the Aug 2025 GPAI-provider wave — the doc gets this distinction right.
• File/path references resolve where they should. packages/arcanea-mcp and apps/web exist in this repo; ARCANEA_PLATFORM_STRATEGY.md (the doc this supersedes) exists at the repo root and does contain the two-domain/monetization content being referenced.
• The doc is upfront about its own limits — the "Provenance" caveat on the Article 50 date and the note to get a one-hour counsel review before the deadline is good practice for a non-legal-advice document.
• No secrets/credentials in the file.

Things worth a second look

1. Cross-repo claims are unverifiable from this repo. Several load-bearing references (CREDIT_PACKS, arcanea-registry-mcp, publishing-house-mcp, world-sdk) don't exist anywhere in this repo — per the doc's own header they live in arcanea-ai-app/Starlight-Intelligence-System. That's disclosed, but since this repo's reviewers can't verify them, consider linking the specific commit/path in the sibling repo where each was checked, so the claims aren't just trust-me references six months from now.

2. "Luminor" as a pricing-tier name may collide with existing terminology. Per .claude/CLAUDE.md and canon, "Luminor" already means (a) the top magic rank (9-10 gates open) and (b) the AI companion entities themselves ("Luminors are declared intelligences," "full Luminor council" as a feature in the Mage tier). Naming the top-paid-tier "Luminor" too means the doc uses the word as a rank, a companion-class noun, and a SKU name in the same table. Might be worth a distinct tier name (or explicitly calling out the intentional overlap) to avoid confusing support/marketing copy later.

3. Pricing ladder skips a canonical rank. CLAUDE.md's Magic Ranks table is Apprentice(0-2) → Mage(3-4) → Master(5-6) → Archmage(7-8) → Luminor(9-10) — five rungs. The pricing skeleton only uses four (Apprentice/Mage/Master/Luminor), dropping Archmage. Not wrong, but if "the pricing ladder is the magic system" (per the closing line), it's currently a compressed/relabeled version of canon rather than a 1:1 mapping — worth a one-line note on why, or align the tier count.

4. Nothing pins this doc against re-drift. Given the doc explicitly says "supersedes the monetization sections of ARCANEA_PLATFORM_STRATEGY.md," it might be worth adding a short note in that older file (or a pointer at its monetization section) back to this one, so a future reader landing on the Feb 2026 doc doesn't act on stale pricing/domain-monetization info. Right now the supersession is one-directional (new doc references old, old doc doesn't point forward).

5. Minor: the "Also" paragraph in section 2 packs marketplace IP terms, MiCA scope, and the payments doctrine ("no autonomous money movement, ever") into one dense paragraph after a numbered list — could be its own subsection or bullet list for scanability, purely a readability nit.

Verdict

Solid, well-caveated strategy doc — legal dates check out, in-repo references check out, no secrets. The two things I'd actually want addressed before merging are #2 (Luminor tier-name collision) and #4 (one-directional supersession pointer), since both are cheap to fix and prevent future confusion; #1 and #3 are lower priority context notes.

[claude]

Review — docs(strategy): Platform & Business Strategy 2026-07

This is a documentation-only PR (a new decision doc + a two-line pointer added to the existing ARCANEA_PLATFORM_STRATEGY.md). No application code, so most of the standard checklist doesn't apply — no runtime behavior, no tests to run, no build to break. Notes below focus on doc accuracy/consistency instead.

Strengths

• The doc is unusually self-auditing: it flags its own weak points inline (the Luminor tier-name collision with the canon magic rank in §3's naming note, and the "not legal advice, get counsel" caveat on the AI Act date in §2). That's exactly the right instinct for a decision doc that will get acted on.
• The supersession note added to ARCANEA_PLATFORM_STRATEGY.md:6 is a clean, minimal edit — a pointer forward rather than rewriting/duplicating content, and the relative link resolves correctly (docs/strategy/ARCANEA_PLATFORM_BUSINESS_STRATEGY_2026-07.md from repo root).
• .arcanea/lore/CANON_LOCKED.md, referenced in §5, does exist in this repo, and packages/arcanea-mcp (also §5/§1) is real here too — the in-repo claims check out.

One accuracy concern worth fixing

• Several file/package paths are stated as fact without distinguishing which repo they live in: arcanea-registry-mcp, publishing-house-mcp, CREDIT_PACKS, and apps/web/app/api/stripe/* (§1, §3, §5, and the PR description) do not exist in this repo (frankxai/arcanea) — only packages/arcanea-mcp does. The doc's header (line 3) does disclaim that "cross-repo claims ... were verified in arcanea-ai-app at drafting time," but that caveat is stated once at the top and then the specific paths are repeated as bare backtick-code throughout without a repo qualifier. Anyone reading this doc in the arcanea repo (which is where it lives) and searching for CREDIT_PACKS or apps/web/app/api/stripe will not find them and may reasonably conclude the doc is wrong rather than realizing it's describing a sibling repo. Suggest either prefixing those specific mentions with "(in arcanea-ai-app)" at point of use, or pulling them into a short "cross-repo assets" callout box so the repo-local vs. cross-repo claims aren't intermixed in the same tables.

Minor / non-blocking

• The Article 50 / Aug 2, 2026 EU AI Act date and "no high-risk category applies" conclusion are load-bearing for the Q3 OKRs (O2) — the doc already flags these need a one-hour counsel review before acting; worth actually tracking that as a checklist item somewhere (e.g., a linked issue) rather than only as prose, so it doesn't get lost.
• The pricing-tier naming conflict with the canonical Magic Rank ladder (Apprentice → Mage → Master → Archmage → Luminor in .claude/CLAUDE.md, vs. this doc's compressed 4-tier Apprentice/Mage/Master/Luminor) is already called out with a fallback plan — no action needed now, just flagging that it's the kind of canon-consistency question the arcanea-canon skill is built to catch, worth a pass before the pricing page ships.

No security, performance, or test-coverage issues — nothing here touches code, secrets, or executable paths.

[claude]

Review

This is a docs-only PR (+180/-0, one new strategy doc + a two-line "superseded" banner on the old one), so the usual code-quality/security/test-coverage lenses mostly don't apply — reviewed for accuracy, internal consistency, and repo-fit instead.

Verified against repo state:

• packages/arcanea-mcp exists as claimed.
• .arcanea/lore/CANON_LOCKED.md exists, matching the canon reference.
• The relative link from ARCANEA_PLATFORM_STRATEGY.md (repo root) to docs/strategy/ARCANEA_PLATFORM_BUSINESS_STRATEGY_2026-07.md resolves correctly.

Strengths:

• Good provenance hygiene: the doc explicitly flags which claims were "verified in arcanea-ai-app at drafting time" vs. this repo, and includes a "re-check before acting on file-level claims" caveat. That's the right posture for a cross-repo decision doc that can't be fully verified in CI here.
• The EU AI Act Article 50 read is explicitly marked as "not external legal advice" with an owner (O2) and a one-hour counsel-review action item before the Aug 2 deadline — good, since that date is load-bearing for the 30/60/90 plan.
• Self-aware naming conflict: the doc flags that the "Luminor" pricing tier overlaps with the canonical top magic rank and the AI-companion entity name, and gives a concrete fallback (rename to "Archmage") rather than leaving it as a latent confusion. Worth flagging per .claude/CLAUDE.md's canon-alignment checklist, but the doc already does the right thing here rather than silently colliding with canon.
• The "supersedes" banner on the old strategy doc is additive (2 lines, no deletions) — preserves history instead of rewriting it, and scopes the supersession narrowly to monetization sections rather than the whole doc.

Minor suggestions (non-blocking):

1. The four-tier pricing ladder (Apprentice/Mage/Master/Luminor) deliberately skips canon's Archmage rank (7–8 gates). That's called out inline, but since .claude/CLAUDE.md treats the Ten Gates / Magic Ranks table as canonical, it might be worth a one-line cross-reference to CANON_LOCKED.md itself (not just the naming note) so a future canon audit doesn't flag it as an unintentional drift.
2. Several claims (CREDIT_PACKS, Stripe routes, publishing-house-mcp) live entirely in the sibling arcanea-ai-app repo and can't be checked from this repo/CI — the doc already says as much, but since this doc will likely outlive the "verified 2026-07-13" freshness window, consider a follow-up issue/reminder to re-verify those cross-repo claims before the Q3 OKRs are treated as committed.
3. Nit: "Days 0–30" in the 30/60/90 plan bakes in an implicit start date (presumably PR-merge date ≈ 2026-07-13) — could be made explicit (e.g., "Days 0–30 (Jul 13 – Aug 12)") so the Aug 2 compliance deadline's position in the timeline is unambiguous at a glance.

No bugs, security issues, or broken links found. Straightforward to merge as a decision record.

[claude]

Review

This PR is documentation-only (a new docs/strategy/... decision doc plus a one-line "superseded" pointer added to the old ARCANEA_PLATFORM_STRATEGY.md), so most of the standard checklist (code quality, performance, security, test coverage) doesn't apply — no code, tests, or runtime surface is touched. Reviewed for factual accuracy, internal consistency, and canon alignment per .claude/CLAUDE.md.

Findings

Canon contradiction on "Luminor" (worth fixing before this ships as a decision doc)
Line 100 states: "Luminor" intentionally overlaps with its canon meanings (top magic rank; the AI companion entities).

.arcanea/lore/CANON_LOCKED.md (the canonical source of truth this doc itself cites) says explicitly:

LOCKED TRUTH: Luminor is a RANK (highest attainment), not an entity type.

So the doc's own naming-conflict analysis is built on a premise canon directly rules out. This doesn't change the decision (four pricing tiers, Archmage skipped), but the stated rationale for the potential "Luminor tier vs. Luminor companion" naming collision is inaccurate — there's no canonical "Luminor as entity" meaning to collide with. Worth a one-line correction so this doesn't get cited later as if it were canon-consistent.

Good practices observed

• Verified packages/arcanea-mcp exists in this repo and is well-documented elsewhere (ARCANEA_USAGE_GUIDE.md, SPRINT_PLAN_WEEK_2026-01-19.md), so that claim checks out.
• The doc is careful to caveat cross-repo claims (Stripe routes, CREDIT_PACKS, publishing-house-mcp, etc. live in arcanea-ai-app, not this repo) and includes a "last verified" / re-check note — good practice for a decision doc whose facts will decay.
• The "supersedes" banner added to ARCANEA_PLATFORM_STRATEGY.md correctly scopes the supersession to monetization sections only, and the relative link path is correct.
• Legal claims (EU AI Act Art. 50, Aug 2 2026 date) are explicitly flagged as "internal read, not legal advice" with an owner and follow-up action — appropriate hedging for a compliance-adjacent claim.

Minor

• The "Luminor" rank-vs-tier ambiguity discussed above is also just a naming risk worth resolving in beta regardless of the canon point — consider deciding it now rather than deferring, since it's cheap to rename before public pricing ships and more expensive after.

Nothing here blocks merge — the canon point is a correction to the doc's own reasoning, not the underlying business decision.

[claude]

Review: docs(strategy) Platform & Business Strategy 2026-07

This is a documentation-only PR (180 additions, 0 deletions, 2 files: a new strategy doc plus a 2-line pointer added to the existing Feb 2026 strategy doc). The usual code-review lenses (bugs, performance, test coverage) don't map directly onto a decision doc, so this review focuses on internal consistency, canon alignment, and reference integrity, plus a couple of content-accuracy notes.

Verified as correct:

• Canon alignment: the pricing tier table (Apprentice, Mage, Master, Luminor) matches .arcanea/lore/CANON_LOCKED.md's locked Magic Ranks table, which has 5 ranks (Apprentice, Mage, Master, Archmage, Luminor). The doc is self-aware about skipping Archmage and explicitly flags it as intentional, not canon drift. It also proactively flags the Luminor rank-vs-entity naming tension that exists elsewhere (e.g. the Luminor-Oracle agent and luminor skill, which use Luminor for companion entities). That is the right way to surface a known inconsistency rather than silently resolving it inside a strategy doc.
• File references resolve: docs/strategy/ARCANEA_PLATFORM_BUSINESS_STRATEGY_2026-07.md (the link target added to ARCANEA_PLATFORM_STRATEGY.md), .arcanea/lore/CANON_LOCKED.md, and packages/arcanea-mcp all exist in this repo.
• Cross-repo scoping is explicit: the doc is careful to call out that apps/web, CREDIT_PACKS, Stripe/Polar routes, and publishing-house-mcp live in the sibling arcanea-ai-app repo, not this one, avoiding the common mistake of implying paths exist locally when they don't.
• The superseding pointer added to the old doc is minimal and scoped correctly (monetization sections superseded; two-domain architecture still stands) rather than blanket-invalidating the older document.

Minor suggestions (non-blocking):

1. Legal/compliance claims could be labeled more consistently. Section 2 includes a good disclaimer (an internal read of the AI Act timeline, not external legal advice, counsel review recommended), but the Aug 2 2026 Article 50 date and MiCA/GDPR/DSA characterizations repeat elsewhere (TL;DR table, 30/60/90 plan) without the same caveat. A single disclaimer near the top covering all legal claims would prevent the TL;DR table being quoted out of context as vetted legal advice.
2. Numbers are unsourced. Pricing tiers, fee comparisons, and OKR targets are stated as decisions/estimates without citation. Since the doc claims to be grounded in and verified against actual repo state, it may help to visually distinguish verified-from-repo facts from directorial estimates/targets.
3. The doc has a short, self-declared freshness window (last verified 2026-07-13, re-verify cross-repo claims again at day 61-90). Good practice, just worth noting this doc's factual claims (not its decisions) may age quickly if referenced later without that caveat attached.

Not applicable to this PR: code quality/best-practices, performance, security, and test coverage checks don't apply, since there is no source code, config, or executable content in this diff, only prose and markdown tables.

Verdict: Clean, well-scoped documentation change. Content checks out against the repo's locked canon and existing file structure, and the doc is unusually careful about disclaiming its own scope (cross-repo paths, legal advice, freshness). No blocking issues.

[frankxai]

PR hospital triage 2026-08-07 — PARK (draft docs/strategy)

Disposition: leave draft. Strategy docs — human strategy gate (docs-ish but not pure public README cleanup).

• Label suggestion: park · ship-candidate-docs · needs-human-gate

[claude]

Review

This is a documentation-only PR (178 new lines in a strategy doc, a 2-line pointer added to the Feb 2026 strategy doc, and 8 .gitignore lines) — so most of the checklist (bugs, performance, test coverage) doesn't really apply here. Reviewing it as a decision document.

What's solid

• The supersession pointer in ARCANEA_PLATFORM_STRATEGY.md:6 is scoped correctly ("monetization sections... superseded", not the whole doc) and the relative link resolves to the right path.
• The doc is self-aware about its own risk: it flags the Art. 50 legal read as "not external legal advice" and explicitly tracks a counsel-review action item (§2, §8) rather than presenting a compliance conclusion as settled fact.
• The pricing-tier/canon tension (skipping Archmage, ranks 7–8) is called out explicitly in §3's naming note with a stated fallback, instead of silently drifting from CANON_LOCKED.md's five-rank ladder (Apprentice → Mage → Master → Archmage → Luminor). Good discipline for a repo where canon consistency is a hard constraint.
• .gitignore additions are clean — none of the newly-ignored paths (.claude-flow/daemon-state.json, daemon.pid, harness-active-policy.json, policy/, .claude/proven-config.json, .claude/.proven-config-version) are currently tracked in the repo, so there's no orphaned-tracked-file issue and nothing to git rm --cached.

Things worth double-checking before this is treated as final

1. Unverifiable cross-repo claims. Nearly every concrete claim (CREDIT_PACKS shape, existing Stripe routes at apps/web/app/api/stripe/*, arcanea-registry-mcp, publishing-house-mcp) lives in arcanea-ai-app, not this repo. The doc dates itself ("verified... at drafting time," "re-check before acting"), which is the right hedge, but since this repo can't verify those claims, it's worth confirming the arcanea-ai-app paths still match before anyone treats §1's integration plan as actionable.
2. Hard date risk. The Art. 50 compliance date (Aug 2, 2026) is treated as ~3 weeks out at drafting time, but today is Aug 16, 2026 — that deadline has already passed relative to when this PR is landing. Worth confirming the O2 counsel review and C2PA/disclosure work actually shipped, or updating the doc's framing so it doesn't read as forward-looking once merged after the fact.
3. No changes to enforce any of this. This PR is purely descriptive — no code, config, or CI changes implement the Polar integration, C2PA labeling, or disclosure requirements it prescribes. That's expected for a decision doc, but the "30/60/90" plan in §8 isn't tracked anywhere else (no linked issue/OKR doc) — a follow-up issue referencing O1–O4 would make those OKRs falsifiable rather than purely narrative.
4. Minor: the pricing table in §3 (€19/€49/€149 tiers) is appropriately hedged as "tune numbers in beta," but since this is the only place actual price points appear in the repo, it'd be easy for someone to later cite this doc as committed pricing without re-reading the hedge.

Security/secrets: none — no credentials, tokens, or .env content in the diff. The .gitignore additions correctly keep local daemon/runtime state and machine-local config out of the repo, which is good hygiene rather than a gap.

Overall: well-hedged, internally consistent with existing canon and the Feb 2026 strategy doc, and honest about its own unverifiable claims. The main actionable follow-up is confirming whether the Aug 2 compliance items actually landed, since that date has now passed.

[claude]

Review

This is a docs-only change (3 files, +190/-0): a .gitignore addition, a superseded-notice pointer added to ARCANEA_PLATFORM_STRATEGY.md, and a new strategy doc at docs/strategy/ARCANEA_PLATFORM_BUSINESS_STRATEGY_2026-07.md. No source code, so most of the usual review categories (bugs, performance, security, test coverage) don't really apply — commenting on what does.

Verified

• Cross-references resolve. The relative link from ARCANEA_PLATFORM_STRATEGY.md to docs/strategy/ARCANEA_PLATFORM_BUSINESS_STRATEGY_2026-07.md is correct, and the new doc's reference to .arcanea/lore/CANON_LOCKED.md exists.
• Canon claim checks out. The doc's footnote asserts "Luminor" is a rank, not an entity type, per locked canon — this matches CANON_LOCKED.md:82 (LOCKED TRUTH: Luminor is a RANK (highest attainment), not an entity type) verbatim. Good diligence given .claude/CLAUDE.md's canon-alignment requirement.
• .gitignore addition is safe. None of the newly-ignored paths (.claude-flow/daemon-state.json, .claude-flow/policy/, .claude/proven-config.json, etc.) are currently tracked in git, so this won't silently un-track committed files — it's purely preventive for machine-local runtime state, which is the right call.

Notes / non-blocking

• Self-aware staleness. The doc is dated 2026-07-13 but merges 2026-08-16, and it candidly flags this itself via the "Merge-time status note" — including that the EU AI Act Article 50 date (2026-08-02) has now passed and the compliance items in §2/O2 are "a live obligation... unshipped as of this merge." That's good practice (dated docs are honest about drift), but worth confirming this doesn't just quietly sit as an overdue item — is there a tracked follow-up issue for the Art. 50 counsel review and C2PA/disclosure work, or does this doc's callout stand in as the only tracking?
• Broad, high-stakes claims for a "decision document." Sections like §1 (Polar.sh as MoR), §2 (legal/compliance posture), and §3 (pricing model) read as settled decisions but rest on assumptions explicitly flagged as unverified by legal counsel (see the provenance note under §2). Might be worth a status marker in the doc header (e.g. "Draft — pending legal review") so it isn't mistaken for ratified policy by someone skimming later.
• Superseded-notice placement. The new pointer at the top of ARCANEA_PLATFORM_STRATEGY.md only calls out "monetization sections" as superseded — reads clean and doesn't over-claim, good scoping.

Nothing blocking. Nice attention to canon/provenance discipline throughout.