docs(common-infra): writing GATE step landed — close last seam + sharpen D1#15
Merged
Merged
Conversation
…sharpen D1 writing's GATE step shipped + tested (writing_gate_probe.py). Every seam is now confirmed across THREE instances spanning both trust-classes (exit-code + semantic) and both compile-outputs (codegen + data). No seam remains design-gated; only writing's A/B parity (a validation step) stands before pass #9. Two substantive updates from writing: - D1 SHARPENED: the runner-side probe.pass is ALWAYS deterministic (exit-code OR a necessary mechanical floor), NEVER a returned judgment. For semantic domains the floor is necessary-not-sufficient and the SEMANTIC AUTHORITY lives entirely OUTSIDE run.js (adversarial review, doctrine #4). So "judgment trust-class" = deterministic-floor-in-runner + semantic-authority-outside — there is no LLM judge inside the probe to game (kills the "haiku judging prose" failure mode by construction). Cleaner than the prior "pass may be a judgment." - Doctrine #6 now has a WORKED INSTANCE: writing closed the emitter gap FIRST (born-canonical OUTLINE from writing-setup + per-section source-pinning from writing-outline) — tolerance eliminated, not relocated. ds/dev still owe this. Extraction readiness: all seams firm; D1's body stays an injected interface (now with all three trust-class shapes as reference) because its computation is intrinsically per-domain, not because it's under-determined. Cleanup items (own recheck channel + durable trigger; pull D4 tier out) noted. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Nv4gnCEvYdR5CqJ9HL3vjA
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
writing's GATE step shipped + tested. Every seam now confirmed across 3 instances (both trust-classes, both compile-outputs); no seam remains design-gated — only writing's A/B parity (validation) before pass #9.
probe.passis ALWAYS deterministic (exit-code or a necessary mechanical floor), never a returned judgment; the semantic authority lives OUTSIDE run.js. Kills the 'haiku judging prose' failure mode by construction.🤖 Generated with Claude Code