Files
zeekayandHanzo Dev fad96e267f no hogs: the root config, the tooling, and the receiver baseline
.stamphog/ was a config directory at the repo root — policy.yml,
review-guidance.md — read by the PR review agent. It is .stamp/ now, and the 34
files that name it follow: the workflows, tools/pr-approval-agent, tach.toml,
the sandbox Dockerfile, bin/debrand.

One behaviour change worth naming rather than burying: the agent finds its own
previous PR comment by an HTML marker, `<!-- stamphog:v1 ... -->`, and that
marker is now `<!-- stamp:v1 ... -->`. Open PRs will get one new comment instead
of an edit, once, and then match again.

insights/test/setup_receivers_baseline.txt records registered signal receivers
by module path, so it names products.review.backend now.

stamphog: zero occurrences anywhere in the tree. review_hog: only inside
migration files, whose names are their identity in django_migrations.

Co-authored-by: Hanzo Dev <[email protected]>
2026-08-09 10:21:41 -07:00

13 lines
550 B
Python

"""Stamp release version.
Stamped onto the review-completed analytics event, the LLM trace properties,
and the verdict comment's mechanics table, so verdict quality and reviewer
behavior can be segmented by version in LLM analytics. Bump it in the same PR
as any behavior-affecting change to the engine, the prompt scaffold, or the
review guidance (semver, pre-releases like 2.0.0b1 welcome). Policy data
changes don't need a bump - they are tracked by the policy sha shown next to
the version in the verdict table.
"""
STAMPFN_VERSION = "2.0.0b4"