[oddharsh]

A copy pass over 19 /garage pages for the tics in declauding 0.13.0. Its linter, run with --skip-quoted on the changed pages, counts 123 hits in the staging families (significance, agency, flat certainty, totalizing, locator, announce, verification, exhaustive) before and 79 after.

What went:

• Grading adverbs with nothing behind them: "genuinely good", "honestly measured", "says plainly", "structurally cannot" (used 4 times on horizon and encoding where the next clause already gives the reason).
• Announced content: "worth saying plainly:", "One thing to notice before moving on.", "Here is the part that made this cheap.", "Worth stating the limit of that probe."
• Significance designations: "That is the part clip-path never did", "the real question", "the interesting question was", "remain the actual answer".
• Constructions reused across pages: "earns its keep/place" (10 times), "the whole idea/problem/trick" (resample and chunks both said "Here is the whole idea"), and "load-bearing" where it meant "required".

Kept on purpose: a quietly or outright sitting beside the contrast it names ("quietly dropped while everything around it reports success"), "byte-identical" next to a measured diff, the juridical "refuses" where a program actually rejects input, quiz distractors, and blueprint's "load-bearing or scar tissue" framing.

Two edits narrow a claim rather than restate it:

• horizon, revert-rule: "nobody knew it had finished" is now "I hadn't noticed it had finished".
• wire, the tripwires: "structurally impossible" is now bounded to the three failures the tripwires check.

Six pages (hidden-flags, octane, pqc, resample, typed-config, useragent) are generator output, so their edits are in pipelines/garage/specs/*.json and the HTML is regenerated. pages:check caught my first hand edit there. check:fast (1777 pass), pages:check, derive:check, format:check and build pass locally. The headers the linter flags (90 comma-clause, 66 coy) are untouched; most are plain labels.

[github-actions]

Wire-size diff vs merge base

62d1987d5a2d293d360063a2cc5162cad8d3f37b → d6fda9152117053d3a59ffffd44f1b9926d5f375

Brotli q11, the quality the edge serves. Every number here is deterministic:
an unchanged file produces no row. Advisory — this check fails on nothing.

Client assets

Total Brotli: 245.02 KiB → 245.02 KiB (-0.00 KiB, -0.0%) across 40 files, 1 changed.

file	base	head	Δ
quiz-data.json	91.23 KiB	91.22 KiB	-0.00 KiB (-0.0%)

Pages

Total Brotli: 720.47 KiB → 720.36 KiB (-0.11 KiB, -0.0%) across 77 files, 17 changed.

No single file moved by 0.13 KiB or more.

_{Plus 17 files under the 0.13 KiB floor, net -0.11 KiB — usually content-hash churn in /a/ references, counted in the total above.}

Worker bundle

No change. 299.54 KiB gzip.

Dictionary transport

Family dictionary acquisition: no change, 14.50 KiB q11.

pd/: 459 → 478 deltas, 542.46 KiB → 549.49 KiB (+7.03 KiB).

Head tags vs merge base

No change. 77 documents, every title, description, canonical, robots, og:*, twitter:*, alternate, JSON-LD type and oracle status identical.

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Preview URL	Updated (UTC)
✅ Deployment successful! View logs	aadhar-sh	db76734	Commit Preview URL Branch Preview URL	Nov 11 2026, 02:22 AM