[codex] Release v0.7.0

[DavidJBianco]

Summary

• Bumps EvidenceForge from 0.6.3 to 0.7.0 for the current dev -> main release.
• Includes the merged slow-test CI split so normal coverage remains on the default suite while slow comprehensive tests run separately without coverage instrumentation.
• Stabilizes the slow release gate by skipping the non-gating 500MB tracemalloc ceiling test and fixing observation manifest collection windows for scenarios that use explicit end times.
• Packages the accumulated dev branch realism, observation profile, source identity, timing, and sidecar-generation work summarized in CHANGELOG.md.

Validation

• uv run ruff check .
• uv run ruff format --check .
• uv run pytest --cov-report=xml (3030 passed, 37 skipped, 79.82% coverage)
• uv run pytest --include-slow -m slow --no-cov --durations=20 (13 passed, 1 skipped, 1:08)

Notes

The branch contains feat: commits since v0.6.3, so the pre-1.0 semver policy calls for a minor bump to 0.7.0.

[DavidJBianco]

Release PR refresh after PR #162 merged into dev:

• dev now includes PR [codex] Clean calibration eval warnings #162 via merge commit 87ac753, so the calibration-cleanup work is part of this dev -> main release PR.
• Version is already bumped to 0.7.0 in pyproject.toml, src/evidenceforge/__init__.py, and uv.lock.
• Added commit 40d17e1 to update CHANGELOG.md and TODO.md with the PR [codex] Clean calibration eval warnings #162 cleanup summary and validation (3075 passed, 15 skipped from that PR).

[DavidJBianco]

Additional CI update landed on dev in 7444042:

• Default/local and feature-PR test lanes now run uv run pytest --no-cov.
• The explicit coverage gate now runs only for PRs targeting main with uv run pytest --cov=evidenceforge --cov-report=term-missing --cov-report=xml --cov-fail-under=70.
• Slow comprehensive tests also remain scoped to PRs targeting main and run with --no-cov.
• Removed default pytest coverage addopts so local/default runs stay fast unless coverage is requested explicitly.

Validation for this CI/docs change:

• Workflow YAML parsed successfully.
• uv run ruff check .
• uv run ruff format --check .
• uv run pytest --no-cov -> 3054 passed, 37 skipped, 34.74s

The release PR body’s older validation section still describes the previous coverage split; this comment records the newer coverage-lane behavior now included in the PR head.