Credibility pass: authorship, show-the-product, and no hand-typed numbers - #2
Merged
Merged
Conversation
A reader had to install and boot ROCC before seeing anything it produces, and the synthetic-by-design rule — the whole governance premise — was below the title rather than in the first sentence. Add two screenshots taken from a local run of this commit's code (Streamlit on a loopback port, captured headless with Playwright, server killed after): docs/assets/rocc-home.png is the default Executive Command Brief; docs/assets/rocc-pipeline-health.png is the aggregate funnel that makes the applicant boundary checkable. Every figure in both images is computed at run time by the deterministic generator, so they reproduce from the pinned seed. The lede now opens on "synthetic by design" and states plainly that there is no hosted deployment. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
CHANGELOG.md stopped at 1.0.0 while twenty-two commits landed behind it, including behavior changes (read-only draft fields, an HTML-escaped heading, a disclosed Hq = Hd forecast assumption). A frozen changelog next to a moving main branch reads as an unmaintained claim. State the rule at the top — this file tracks releases, not commits, and git log is the authority — then list the post-1.0.0 work under Unreleased, derived from the actual git history rather than memory. The version strings stay at 1.0.0 because no newer version has been cut, and the header now says so instead of leaving a reader to infer it. Also name the gate commands in the README and record that CI provisions Python 3.11. Every number the README and CHANGELOG state about this repo was re-checked against the code first: the nine-page count matches PAGE_RENDERERS, the three named contact surfaces match ADR-024, the six-horizon fan matches FORECAST_HORIZONS, and byte-stable generation was confirmed by running the generator twice and comparing SHA-256 sums. "Seven dead-code items" from a commit subject could not be pinned to exactly seven items, so the changelog entry says "unused imports and locals" instead of carrying a number forward. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The repo was silent on authorship while its history is full of AI co-author trailers. Silence reads worse than the truth, and it hides the part of the work that is actually mine. State the division of labor plainly, name the gate every slice had to pass and the adversarial review that ran before merge, and give one concrete catch rather than a claim of rigor: the role-gated applicant rows rejected in ADR-024, including the reasoning — a role dropdown in a demo UI is a display preference, not an authorization boundary, so gating rows by role protects nobody. The fix was deleting the gate, not tuning it. Placed after the substance and before the license, and kept to prose. Nothing in it is a claim a reader cannot check against ADR-024, the CI workflow, or the commit log. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
…build note Three corrections found by an adversarial read of the previous docs pass. 1. The home-screenshot caption claimed Site Readiness and Ratio Forecast were "the other two surfaces that project a ratio." Reports is a fourth: pages.py:888 pulls the same forecasts, and the reports it renders print a projected ratio (services.py:137 for the leadership summary, services.py:101 for the site report). Reports carries DRAFT_BANNER (pages.py:887), not PLANNING_BANNER, so the old sentence also implied governance coverage that does not hold. The caption now names Reports and says which banner it carries, plus the planning-indicator line the site report repeats in its own body. 2. The lede and rule 2 stated the same governance rule differently 24 lines apart: "aggregate pipeline stages" versus "contracts". The lede now matches rule 2, pages.py:953, and docs/PRODUCT_BLUEPRINT.md:41. Both README statements also move from "get measured" to the code's "may be measured" — the rule is a permission boundary, and contract-level work is on the roadmap as not built. 3. "How this was built" no longer claims a personal catch or an independent review, and no longer says the review pass ran first. Commit 1b7b24c ("writer-review fix") is a direct child of the initial release, so the pass ran after that work landed. The co-author trailers are now credited with only what they evidence: every non-merge commit names an AI co-author. The gate is named as CI runs it, including the container build and health probe. Gate after the change: 23 passed, ruff clean, synthetic validation passed. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
"the one CI still runs on every push" implied continuity the history disproves: ruff entered CI at 84c625c (17 of 26 commits predate it) and the advisory scan and health probe at f5921df (19 of 26 predate them). Split into what the slices cleared and what CI has grown since. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
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.
Items 1, 2 and 7 of a ten-item review punch list.
1 —
## How this was built. Credits the review passes with the finds; scopes the trailer claim to the AI half of the split, which is the half trailers can actually evidence.2 — Show the product. Two local-boot screenshots, and the first screen now states plainly that all data is synthetic by design — a governance strength here, not a limitation.
7 — Numbers. CHANGELOG now says what it tracks and is current.
Honest notes
ruffentered CI at84c625c(17 of 26 commits predate it) and the advisory scan and health probe atf5921df(19 of 26 predate them).DRAFT_BANNERrather thanPLANNING_BANNER. Corrected.Gate: green.