c3537f9dd4
Both atom substrate (6-week lock) and agent substrate (3-week lock) were calibrated for calendar windows that were 30-100× slower than actual execution. Phases shipped in hours, not weeks. Keeping calendar locks delays breaking consolidations that can safely ship now. Unlocks: verb templates, schema fragments, PatternGate, CommandVerify, role expression, DNA identity — all can ship as breaking changes with standard review gate (audit + integration tests). No more calendar-week locks. Future locks: agent-hours or "until declared phases land." Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
53 lines
2.3 KiB
Markdown
53 lines
2.3 KiB
Markdown
# Schema Locks — REVOKED
|
||
|
||
**Revoked on:** 2026-04-23
|
||
**Revoked by:** user (explicit request "Мы проекты за часы делаем — недели abstraction")
|
||
**Replaces:** `docs/SCHEMA-LOCKED.md` and `docs/AGENT-SCHEMA-LOCKED.md`
|
||
|
||
## Revocation rationale
|
||
|
||
Both schema locks (atom substrate v1 and agent substrate v1) were originally declared for calendar windows:
|
||
- Atom substrate: 6 weeks (through 2026-06-03)
|
||
- Agent substrate: 3 weeks (through 2026-05-14)
|
||
|
||
Reality check: the phases those locks were intended to protect shipped in **hours**, not weeks:
|
||
- Atom substrate phases 1-5 (4 parallel streams + integration) — landed in ~2 hours
|
||
- Agent substrate phases 1-5 — landed in ~30 minutes
|
||
- Convergence pre-unlock wave (U1+U2+U3) — landed in ~25 minutes
|
||
|
||
**The locks were calibrated for a timeline that was 30-100× slower than actual execution.** Calendar-week windows are corporate-ritual pacing, not real-velocity pacing. Keeping the locks would delay breaking consolidations (path-filter merge, cargo-green merge, /audit target, verb-template refactor) that can now safely ship.
|
||
|
||
## What revocation enables
|
||
|
||
Immediate (post-revocation, in-flight wave):
|
||
- Verb templates (CRUD as data) — Layer A from convergence plan
|
||
- Schema fragments ($ref'd JSON) — Layer B
|
||
- PatternGate unified trait — Layer C
|
||
- CommandVerify unified trait — Layer D
|
||
- Role expression (extends/adds/relaxes) — Layer E
|
||
- DNA identity — new Layer G (agent ID encodes composition)
|
||
|
||
Follow-up:
|
||
- `scope::path-filter` (consolidate whitelist + denylist)
|
||
- `quality::cargo-green` (consolidate check + tests)
|
||
- `/audit <target>` (collapse 6 skills + checklist registry)
|
||
- kei-runtime-core extraction (shared infrastructure across 3+ runtime crates)
|
||
|
||
## What stays locked (nothing)
|
||
|
||
Neither SUBSTRATE-SCHEMA.md nor AGENT-SUBSTRATE-SCHEMA.md is locked. Breaking changes allowed from this commit forward with standard review gate (audit agents + integration tests pass).
|
||
|
||
## Process change
|
||
|
||
Future schema-lock declarations:
|
||
- Express duration in **agent-hours-of-work**, not calendar weeks
|
||
- Default: "locked until explicitly revoked OR all declared-dependent phases land"
|
||
- Revocation requires: ledger row + reason + this doc updated
|
||
|
||
No more calendar-week locks.
|
||
|
||
## Ledger entry
|
||
|
||
```
|
||
2026-04-23 schema_lock_revoked atom,agent reason=calendar-overestimate,phases-landed
|
||
```
|