Skip to content

test(prose): three cases for the truth stack + three-phase staleness fixes - #960

Open
leeovery wants to merge 1 commit into
discussion-measured-claimsfrom
prose-truth-cases
Open

test(prose): three cases for the truth stack + three-phase staleness fixes#960
leeovery wants to merge 1 commit into
discussion-measured-claimsfrom
prose-truth-cases

Conversation

@leeovery

Copy link
Copy Markdown
Owner

Why

The stack's behaviours were unwalked, and the existing spec cases pinned the old two-phase review — their stub triggers and asserts would have misdirected the next walk.

What

Three new cases, one per shipped behaviour:

Staleness fixes: the spec-review-clean stub and the four cases arming it move to the three-agent trigger; spec-extracts-the-discussion and spec-resolves-a-source-conflict asserts update to the three-phase sequence.

Corpus validation, snapshot goldens (all 71 worlds rebuild byte-identical), conventions lint, and typecheck are green. Walks are on-command via /prose-test as ever.

Stacks on #959. Part 6 of the idea-40 stack.

🤖 Generated with Claude Code

…fixes

New cases, one per shipped behaviour: spec-measures-a-false-claim (the
reality trigger — construction verifies a load-bearing claim, stops
conversationally when the falsity undermines the leaning decision, and
lands the settlement in the discussion's own record),
spec-review-routes-a-source-defect (the review loop's fourth verb — a
claims finding whose false count lives in the source routes on the
value-only lane, corrects the discussion in place, and re-aligns the
spec's copy), and discussion-review-corrects-a-false-claim (document
review's unverified-claims dimension re-runs the recorded command at
conclusion and corrects the value without turning arithmetic into a
decision). A new spec-claims-source-defect stub carries the findings
return; the spec-review-clean stub and the four cases arming it move to
the three-agent trigger, and the two asserts pinning the two-phase
sequence update.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
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.

1 participant