[kanfil]

Release cut per factory-deploy publish step (order: merge → tag → changelog).

• Tag v2.7.0 already pushed (tag-as-truth); this PR carries the version bump + CHANGELOG onto main.
• Merges: feat(team-setup): deterministic --directives wire + non-TTY fail-loud guard #65 (7d6d502) + fix(opencode): route session_end through the session.deleted lifecycle event #66 (48353fc). CI: 10/10 on both PRs, local 838/838 on release commit.
• GitHub release: https://github.com/tikalk/adlc-cli/releases/tag/v2.7.0

[kanfil]

factory-review verdict: merge-ready (PR #67)

Reviewed revision: head=eb9e2bdf4deae6c167159e9091b96349c8e521df, base=48353fce019d09d3dcb47882ef480c94fc027bf6, merge-base=48353fce019d09d3dcb47882ef480c94fc027bf6 (reviewed from detached checkout .adlc/worktrees/factory-review-eb9e2bd, not the working tree).

Scope: 2 files, 21 insertions(+), 1 deletion(-) — package.json version bump 2.6.0 → 2.7.0 + CHANGELOG.md [2.7.0] section. No source, test, workflow, or CI-config changes.

Pass results (REVIEW.md)

Pass	Severity	Result
bugs (null/undefined, unhandled errors, resource leaks)	Important	✅ PASS — no executable code; package.json parses as valid JSON
security (injection, secret exposure, auth bypass)	Important	✅ PASS — new CHANGELOG text contains no secret-like strings (grep hits are pre-existing historical entries); no auth/CI surface touched
spec-compliance (PRD/AD coverage)	Important	✅ PASS — release-cut docs only; repo has no docs/adlc/product/PRD.md / docs/adlc/architect/AD.md to cover; notes accurately describe #65 (Added) and #66 (Fixed)
quality (naming, dead code, docs)	Nit	✅ PASS — Keep-a-Changelog format consistent with prior entries; version/date header well-formed

Cross-checks

• Tag/version consistency: tag v2.7.0 exists and package.json at reviewed head reads 2.7.0 — matches publish.yml tag-gate (TAG == PKG).
• CI: all test matrix jobs pass (18/18 across ubuntu/macos/windows × 18/20/22) + e2e-skip-agent pass. The single publish failure belongs to the tag-push Publish workflow run (my client-side gh release create collided with the workflow's own gh release create → HTTP 422 tag_name already exists) — npm publish itself succeeded (+ adlc-cli@2.7.0). Not a PR defect; not in the PR's check rollup path.
• Review gate: reviewDecision: REVIEW_REQUIRED — human code-owner approval still required; this review is COMMENT-type only and does not approve.

Findings: 0 Important, 0 Nit. Verdict: merge-ready — awaiting human code-owner sign-off.