Source-28 governance consolidation
Replaced competing living-root governance with a 28-source canonical set and separate master handoff.
Paused by design. Authority sealed. Next gate waiting.
A public engineering record of what is proven, what is parked, what broke, what survived, and exactly where the build resumes.
Verified from the accepted Source-28 seal evidence. The page itself is a side quest and does not advance product authority.
A modular, multi-tenant, multi-user, role-aware, API-first, agent-native business operating platform whose first commercial wedge is CRM.
PRAXIS keeps failed gates visible. A STOP is evidence, not something to rewrite out of the story.
Earlier foundation phases are closed/sealed historical authority. The current Source-28 handoff preserves that history but does not enumerate every earlier sub-gate.
Truth census completed and evidence recorded.
Contract candidate passed and was sealed.
Authorization candidate passed 34/34 tests and was sealed.
Data/migration contract candidate passed and was sealed.
PostgreSQL 18 tenancy/governance candidate passed 27/27 assertions and was sealed.
Runtime authorization candidate was not accepted. The stop is preserved as evidence.
Runner/package remains parked and explicitly unauthorized.
R5 installed the exact 28-source governance canon plus master handoff without runtime, DB, provider, Git-history or parked-P4F mutation.
29 governance paths committed and tagged; parked P4F hashes remained unchanged.
PRAXIS intentionally paused here for a temporary public status-page side quest.
The Source-28 side quest exposed multiple PowerShell/harness failure modes. None were allowed to silently become product authority.
| Attempt | Result | Problem | Containment | Correction |
|---|---|---|---|---|
| Initial Source-28 installer | STOP | Call-depth overflow after package-manifest validation. | No acceptance claimed; installer abandoned. | Rebuilt as a standalone non-recursive runner. |
| Source-28 R3 | STOP | PowerShell 5.1 StrictMode exposed scalar .Count behavior. | Failure occurred before documentation install/write stage. | Replaced scalar-sensitive count checks with PowerShell-5.1-safe enumeration. |
| Source-28 R4 → R5 | RECOVERY | Evidence location was too buried for the operating workflow. | No authority change; evidence workflow corrected. | R5 emitted PASS/STOP evidence directly into Windows Downloads and then passed. |
| Source-28 seal R1 | STOP | PowerShell parser rejected variable interpolation followed by a colon: $hookCode: / $newHead:. | Script never parsed, so no Git mutation executed. | Delimited variables and added parser preflight. |
| Source-28 seal R2 | STOP | A helper parameter named $Args collided with PowerShell's automatic $args variable, causing a bare git invocation. | Stopped at first Git state read; no commit or tag created. | Rebuilt the Git helper and added argument-forwarding self-test. |
| Source-28 seal R3 | PASS | None remaining at gate execution. | Exact governance-only stage set enforced. | 29 governance paths sealed; parked P4F preserved byte-for-byte. |
These four known P4F candidate paths are intentionally dirty/parked. Dirty does not mean wrong. They must not be reset, cleaned, overwritten, staged or promoted by accident.
These were not random detours. Each one reduced ambiguity, hardened evidence handling, or protected the sealed boundary.
Replaced competing living-root governance with a 28-source canonical set and separate master handoff.
Converted repeated PowerShell/harness failures into parser preflight, argument-forwarding tests, fail-closed evidence and exact mutation boundaries.
Changed runners so PASS/STOP evidence lands directly in Downloads for immediate review.
Temporary public read-only showcase/status surface for the meeting. It does not advance or mutate PRAXIS product authority.
The queue is frozen, not abandoned. Exactly one implementation gate is next when PRAXIS resumes.
| Gate | Foundation slice | Status |
|---|---|---|
| P4F | Actor + membership authorization runtime v2 | NEXT / FROZEN |
| P4G | Capability registry + module-manifest foundation | QUEUED |
| P4H | Products, plans, plan versions and entitlements | QUEUED |
| P4I | Usage meters, budgets and credit ledger | QUEUED |
| P4J | PRAXIS SaaS billing adapter — sandbox first | QUEUED |
| P4K | Release fabric, rollout, dependencies and kill switches | QUEUED |
| P4L | Tenant commerce domain contract | QUEUED |
| P4M | Embedded payments / Stripe Connect foundation — sandbox | QUEUED |
| P4N | Agent identity, delegation and risk | QUEUED |
| P4O | Transactional outbox + durable execution | QUEUED |
| P4P | Memory/cache foundation | QUEUED |
| P4Q | Public API/webhook/service-account/MCP seam | QUEUED |
| P4R | Cross-layer integration/torture/security gate | QUEUED |
| P4S | Commercial platform foundation seal | QUEUED |
Serious CRM for small teams, expanded continuously by sealed capabilities.
The status page does not inflate claims. It reflects the same default-deny, evidence-first posture as the product.
When the side quest ends, PRAXIS returns here. No drift, no stale runner, no accidental cleanup.
This is a static public snapshot of the latest verified PRAXIS evidence, not a direct connection to the private repository or production systems.
Viewer clock: