Repository navigation
test(plugin): every count of the audit's checks is AdCopilot's 70 (stacked on #25) - #28
Closed
SaiSatya16 wants to merge 1 commit into
Closed
SaiSatya16 wants to merge 1 commit into
SaiSatya16 wants to merge 1 commit into
Conversation
AdCopilot v2.44 lists 70 checks (D221) and holds every count it states to that number with one test (tests/test_check_count_agreement.py, D225). This is the same check for the plugin's own text: every tracked file but tests/, the README's release notes left as history. A count of the checks above 70, a seventy-three or a mock's checks_total that is not 70 fails, by file and line. A count of 70 or under may be a run's own and is left alone. On main (before #25) it names the twelve stale counts #25 rewrites; on this branch it passes. 0 Google operations, no version bump, no release.
SaiSatya16
pushed a commit
that referenced
this pull request
Oct 22, 2026
PR #28 (audit/checks-info @ db103b5, stacked on #25) squashed in: tests/test_check_count.py. Its scan leaves out the stamped eval tool snapshots (*/mocks/adcopilot/_tools.json), which keep the count they were taken with until plugin issue #5's whole refresh from the server (0.2.15); test_only_stamped_snapshots_are_skipped holds that only a file whose descriptions carry a "Tools revision:" stamp is left out. #25's edits to those snapshots stay out, as in the commit before.
SaiSatya16
pushed a commit
that referenced
this pull request
Oct 22, 2026
…n the order it matters One note for the version, as the release checklist reads it: written against AdCopilot v2.44.0 (deployed 2026-10-22), none of its steps run. - First the reason the version goes out first: the daily set-up lets "AdCopilot: what is connected and where you are" and "AdCopilot: save business notes and progress" through (every other Write/delete tool still on Needs approval), and the recipe's first call names the morning check's task. - Second #22's scheduled marker from every file that sends the plugin's. - What does not change: the daily run still calls no save_org_context; the ban lifts only once the people who already schedule the check have been told to set the two tools to Always allow and that notice has been measured (AdCopilot notice_2026_10_schedule_allow). - Third to eighth #24's six changes, as #24 wrote them; ninth #25/#28's seventy checks; tenth #26's ready calls and drafts. - Release order: every server half is live in v2.44.0. The Evals section counts thirty cases (#22's review-scheduled-unmarked and #24's fix-checks-three-checks both new), and names both as written and never run.
SaiSatya16
added a commit
that referenced
this pull request
Oct 23, 2026
…ts first call (#30) Includes #22, #24, #25, #26 and #28. The daily set-up puts AdCopilot's two note tools ("AdCopilot: what is connected and where you are", "AdCopilot: save business notes and progress") on Always allow beside the read-only tools, so a scheduled run no longer waits at its first call; the recipe's first call names the morning check's task; a scheduled run sends claude-scheduled from every skill; the adcopilot-fix-checks skill; seventy checks; an audit finding's ready call and draft. The daily run still calls no save_org_context. Needs AdCopilot v2.44.0, live since 2026-10-22. Refs AtromxIntelligence/adcopilot#942 (v2.44.0)
Contributor
Author
|
Released in 0.2.14 (#30) |
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
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.
What
One test,
tests/test_check_count.py: every count of the audit's checks in this repo is AdCopilot's 70.It reads every tracked text file except
tests/(the tests quote stale phrases to ban them) and fails, by file and line, on:full_auditmock whosechecks_totalis not 70.A count of 70 or under is left alone, because it can be a run's own ("53 of the 59 checks that apply", "15 checks not evaluated"). The README's release notes are history and keep their counts. One eval comment that quotes what 0.2.11 said is listed in
ALLOWEDwith its reason, and a listed line that disappears fails too.Stacked on #25 (
audit/honest-count-score, the count change itself): base is that branch. Merge #25 first.Why
Spec
_tools/audit-design-final.json, plan.catalogue_design §1: "the public plugin repo (skills/adcopilot-daily/SKILL.md '73 checks', eval mocks '74') gets its own check and PR". #25 is the PR; this is the check. AdCopilot holds every count it states to its registry with the same kind of test (tests/test_check_count_agreement.py, D225 in the AdCopilot decisions log, lane checks-info, AtromxIntelligence/adcopilot#947).Switch, defaults, ops cost
Tests
python3 -m unittest discover -s tests: 215 passed (210 before, plus this file's 5).origin/main(before docs(plugin): AdCopilot v2.44's audit count: 70 checks, each listed, and the score from the checks that apply #25) the check fails and names the twelve stale counts docs(plugin): AdCopilot v2.44's audit count: 70 checks, each listed, and the score from the checks that apply #25 rewrites (the auditor's "seventy-three", the_tools.jsonand replay "74 checks", the mocks'checks_total: 73). On this branch it passes.Committed as Atromx Intelligence support@atromx.com. Do not merge or release this before AdCopilot v2.44.
🤖 Generated with Claude Code