Remove the Codex LLM review and the per-PR image job — both spent API
tokens on every Bossbot run and the connector review already covers
code review. The gate is now fully deterministic: Tests passed for the
head SHA, non-draft, trusted author, zero unresolved review threads.
Prompt/schema files removed; guard tests updated to forbid
openai/codex-action and image generation from reappearing.
Signed-off-by: phernandez <paul@basicmachines.co>
Mirror the ProjectUpdateContext shape the basic-memory.yml capture flow
collects: the image prompt now receives a compact change-shape digest —
labels, linked issues with titles, commit subjects (the PR's narrative
arc), and a churn-ranked changed-files summary with totals — from a
single 'gh pr view --json' call passed as --pr-context-file (replacing
--pr-title/--pr-body-file). The digest is explicitly context, not
captions: the prompt forbids rendering paths, stats, issue numbers, or
commit subjects verbatim in the image.
Signed-off-by: phernandez <paul@basicmachines.co>
Two BM Bossbot fixes:
1. Unresolved review threads now block approval. The review prompt only
sees metadata+diff, so an approve verdict said nothing about
outstanding feedback — #932 merged with two open Codex P2 threads.
finalize now counts unresolved reviewThreads via GraphQL and fails
the status when any remain. A new recheck command, triggered by
pull_request_review / review_comment / review_thread events, flips
the status to failure when feedback arrives after approval and
restores a previously earned approval for the same head SHA once
every thread is resolved. Review and recheck jobs use separate
job-level concurrency groups so neither cancels the other.
2. PR images depict the PR's content, not the review outcome. The image
prompt previously used the BM Bossbot review summary (verdict/status)
as its only source material, so every image rendered an APPROVED
stamp. The prompt now sources the PR title and the author's own
description (managed bot blocks stripped) and explicitly bans
approval stamps, verdict wording, badges, checkmarks, and Bossbot
branding. Theme selection seeds on PR number+title for stability
across re-reviews.
Signed-off-by: phernandez <paul@basicmachines.co>
The coverage collection had multiple issues:
- Wrong pytest markers caused 0 tests to run
- Postgres jobs silently skipped artifact uploads
- Coverage Summary job failed when artifacts were missing
- uv venv picked wrong Python version for coverage jobs
Simplify: every job just runs tests via `just` recipes. No more
dual code paths, artifact uploads, or summary aggregation job.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Signed-off-by: phernandez <paul@basicmachines.co>
The coverage code paths for Postgres unit and integration jobs filtered
with `-m postgres`, but no tests use that marker. Postgres is selected
via BASIC_MEMORY_TEST_POSTGRES=1 env var. This caused 0 tests to run
on Python 3.12 (the only version with coverage: true).
- Postgres unit: remove `-m postgres` (matches `just test-unit-postgres`)
- Postgres integration: use `-m "not semantic"` (matches `just test-int-postgres`)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Signed-off-by: phernandez <paul@basicmachines.co>
Previous commit still ran tests twice on the coverage matrix entry.
Now the 3.12/ubuntu/main combo runs with coverage directly, and all
other matrix entries run without. No duplicate work.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Signed-off-by: phernandez <paul@basicmachines.co>
The coverage summary job was re-running the entire SQLite + Postgres test
suite from scratch (~60 min), duplicating work already done by upstream jobs.
It consistently timed out.
Now each test job collects coverage data and uploads it as an artifact.
The coverage job just downloads, combines, and reports — should take <1 min.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Signed-off-by: phernandez <paul@basicmachines.co>
Add Postgres service container and separate test step for PostgreSQL backend testing.
The Postgres tests only run on Linux runners since GitHub Actions service containers
are only available on Linux.
- Add postgres:17 service container with health checks
- Add 'Run tests (Postgres)' step with Linux-only condition
- Rename existing test step to 'Run tests (SQLite)' for clarity
This enables CI testing of dual database backend support introduced in the
postgres-support feature branch.
Signed-off-by: phernandez <paul@basicmachines.co>
Fixed the conditional logic in claude.yml to properly handle different event types:
- Use github.event.comment.author_association for issue_comment events
- Use github.event.sender.author_association for other events
- Maintain support for all basicmachines-co org members (OWNER/MEMBER/COLLABORATOR)
This ensures @claude mentions in PR comments trigger the workflow correctly.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Signed-off-by: phernandez <paul@basicmachines.co>
phernandez