Skip to content

ground-truth-gates: a gate over hardcoded facts asserts the facts, not just the shape - #79

Merged
F-e-u-e-r merged 1 commit into
F-e-u-e-r:mainfrom
firaen22:mine-chat-evidence-as-fixture
Jul 26, 2026
Merged

ground-truth-gates: a gate over hardcoded facts asserts the facts, not just the shape#79
F-e-u-e-r merged 1 commit into
F-e-u-e-r:mainfrom
firaen22:mine-chat-evidence-as-fixture

Conversation

@firaen22

@firaen22 firaen22 commented Jul 24, 2026

Copy link
Copy Markdown
Contributor

What

New rule 6 under "What makes a gate real". When code embeds domain constants (holiday dates, a tax rate, a fee schedule, a valid-state set), a structure-only suite passes identically whether the values are right or a later edit corrupted one. Those values were usually cross-checked once — but in the conversation, which evaporates at session end. Anchor the load-bearing values as assertions from the authority you already consulted, so a future silent constant change fails.

Why

This extends item 2's carve-out — from an output-interface string to an embedded input constant — and is distinct from item 3's "hardcoded expected value" fake-pass: the anchor's value comes from an external authority, not copied back from the code's own output. Item 3's tautology asserts the source is the source; this asserts the source matches the world.

Class-distilled from a mining pass over the owner's own sessions (no code taken): a "tests check shape, not date-truth" cross-model finding on hardcoded holiday dates that had been cross-checked against four sources — but only in chat, never in the gate.

Ships unprobed per the covenant. .github/checks.py green.

Revision note (2026-07-24)

Force-pushed after a same-day adversarial self-audit found the original "Why" over-claimed: it said this rule "IS item 2's carve-out used correctly," but item 2's carve-out is explicitly scoped to output-interface contracts with downstream consumers (an error-message string, an output name) — a holiday date is an embedded input constant, not an interface. Reframed as an extension of that carve-out rather than a direct instance of it, and added the reconciliation with item 3's "hardcoded expected value" fake-pass (left unaddressed in the original draft) — the two are distinguished by the expected value's provenance, not its presence.

🤖 Generated with Claude Code

@firaen22
firaen22 force-pushed the mine-chat-evidence-as-fixture branch from ea60957 to 1fab936 Compare July 24, 2026 15:52
…t just the shape

New rule 6 under 'What makes a gate real'. When code embeds domain constants
(holiday dates, rates, valid-state sets), a structure-only suite passes
whether the values are right or a later edit corrupted one. The values were
cross-checked once — but in the conversation, which evaporates at session end.
Anchor the load-bearing values as assertions from the authority you already
consulted, so a future silent constant change fails. This is item 2's
change-detector carve-out used correctly (the value IS the contract).
Class-distilled from a mining pass over the owner's own sessions (no code
taken): a 'tests check shape, not date-truth' cross-model finding on hardcoded
holiday dates.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
@firaen22
firaen22 force-pushed the mine-chat-evidence-as-fixture branch from 1fab936 to 5a04f92 Compare July 25, 2026 00:11
@F-e-u-e-r
F-e-u-e-r merged commit 0af0213 into F-e-u-e-r:main Jul 26, 2026
3 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants