docs: align v0.8 adoption release guidance - #534

Claude audit (merge-authority lane)

Head SHA: 4f373823c7b2da65adfbc5710c9ee327a598a25f
Findings: P0=0, P1=0, P2=0, P3=2 (blocker policy: any P0/P1/P2 -> BLOCKED)

Claude Audit: PASS

Summary:

Docs-only consistency pass aligning v0.8 adoption guidance (doctor preset renamed, pipx install pattern hardened, stale beta/version references removed), backed by an updated regression test (test_active_docs_use_current_beta_release_language) that greps the touched docs for stale phrases. Cross-checked the diff against the new/updated test assertions and found the docs match what the tests require; no code paths changed. No P0/P1/P2 issues found.

Findings:

• [P3] Loosened production PyPI publish gate -- docs/pypi-release.md:146
  Finding ID: claude:6345c2bf81fe157f9316
  The dispatch-matrix guidance changed from 'production publish only after TestPyPI passes' to allowing routine beta publishes straight to production PyPI once the no-publish verification run is green, with TestPyPI reserved for 'first-time trusted-publishing setup or risky packaging changes.' This is a real reduction in the pre-production safety net (PyPI releases are effectively irreversible) but it's stated with explicit rationale and is an internally consistent, deliberate documentation choice rather than an objective defect — flagging for the release owner to confirm intent, not a merge blocker.
• [P3] Subject-verb mismatch in updated release-notes prose -- docs/early-adopter-v05.md:157
  Finding ID: claude:b32a079b6839911f521d
  "Later v0.5 beta dogfood follow-ups keep owner-bound work out of lane WIP, adds lane-liveness..." mixes plural 'follow-ups keep' with singular 'adds' in the same sentence — a copy-editing nit, non-blocking.

Code Mower notice: previously audited head 4f373823c7b2 is no longer an ancestor of current head 8791ca7efe3b; commits may have been dropped. Only current-head audit verdicts count.

gitar auto-apply:on         

gitar display:verbose         

_{Was this helpful? React with 👍 / 👎 | Gitar}

Claude audit (merge-authority lane)

Head SHA: 8791ca7efe3be2e6e035c3771310e0d0a3f4b146
Findings: P0=0, P1=0, P2=0, P3=1 (blocker policy: any P0/P1/P2 -> BLOCKED)

Claude Audit: PASS

Summary:

Docs-only PR that aligns v0.8 adoption/release guidance across seven docs and adds a corresponding release-hygiene test (test_active_docs_use_current_beta_release_language) that verifies stale v0.5/v0.6/v0.7 phrasing was removed and current beta language is present. Spot-checked the new test's assertions against the actual doc edits; all match. No code paths, CLI behavior, or CI gates are changed.

Findings:

• [P3] Release runbook loosens TestPyPI-first guidance -- docs/pypi-release.md:178
  Finding ID: claude:90d61311f1c5dd682866
  The matrix row and surrounding prose now say routine beta publishing 'can go from the green no-publish verification run to production PyPI' and only recommends TestPyPI 'for first-time trusted-publishing setup or risky packaging changes,' replacing the prior blanket 'Use only after TestPyPI passes.' This is a real policy relaxation for a security-sensitive publish gate. It doesn't appear to contradict docs/oss-v1-checklist.md (which plausibly covers the one-time initial public release setup rather than routine releases), so it is not clearly a broken contract, but worth a maintainer sanity check that this is the intended risk tradeoff.