Skip to content

insights: close the post-merge review findings on workflow intelligence#763

Merged
iamtoruk merged 1 commit into
mainfrom
fix/756-review-findings
Jul 20, 2026
Merged

insights: close the post-merge review findings on workflow intelligence#763
iamtoruk merged 1 commit into
mainfrom
fix/756-review-findings

Conversation

@iamtoruk

Copy link
Copy Markdown
Member

Closes the five open findings from the in-depth review on #756 (F1 was already fixed pre-merge in f8c9a6a; this handles F2-F6 plus one non-blocking note).

  • F5: corrections count only follow-up prompts (the opening prompt is a task, not a correction) and the revert pattern matches demonstratives only, symmetric with undo. Regression tests: correction-shaped opener not counted, 'revert the ' follow-up not counted.
  • F3: an unparseable timestamp on the FIRST edit yields null for the session instead of measuring to the first parseable edit.
  • F4: file churn normalizes backslashes before keying/relativizing/basenaming; the payload's basename redaction now works on Windows paths (previously the full absolute path, username included, could ship). Test covers mixed-separator dedup and relativization.
  • F6: the pricing-coverage denominator excludes deliberately-free models via a new isExpectedFreeModel that mirrors findUnpricedModels' exclusions.
  • F2: pricingCoverage is number | null on the payload; absent renders as null, never a fabricated 1.0.
  • Note: the one-shot min-edit-turns threshold is now a shared constant between worstOneShotCategory and the coaching gate.

Full suite green (3795 tests). Real-data drive: headline still equals the history sum; corrections dropped from 14 to the follow-up-only count; coverage and reworked files intact.

Five verified findings from the #756 review, each with a regression test:

Corrections now count only follow-up prompts. The session-opening prompt
cannot be correcting this assistant, however correction-shaped its task
text reads, and the revert pattern no longer matches ordinary 'revert
the <noun>' requests (undo was already guarded; the doc claimed both
were).

Time-to-first-edit returns null when the first edit's timestamp is
unparseable instead of silently measuring to the first parseable one.

File churn normalizes backslashed paths before keying, relativizing, and
basenaming. Previously the payload's basename-only privacy redaction was
a no-op on Windows and shipped the full absolute path, username
included; mixed separators also split one file across churn entries.

Pricing coverage excludes deliberately-free models (local-looking,
local-savings mapped, zero-rate overrides) from the denominator via the
same exclusions findUnpricedModels applies, so a mostly-ollama user no
longer reads high coverage while every cost-bearing call is unpriced.
An absent coverage value now reaches the payload as null, never as a
fabricated 100%.

Also shares the one-shot minimum-edit-turns threshold between
worstOneShotCategory and the coaching gate so they cannot drift.
@iamtoruk
iamtoruk merged commit db8ef9b into main Jul 20, 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.

1 participant