[thebrandonlucas]

Stack: master ← #26 ← #27 ← #28 ← #29 ← #30 ← #55 ← #60. #20, #22, #57, #24 and #19 are merged.

• Troubleshooting (docs/guides/troubleshooting.md): an entry per exit code and per credential_problem (no_credential, multiple_credentials, unknown_credential, empty_api_key, bound_elsewhere, auth_url_mismatch), plus expired logins, API-key discovery, HTTP 401/403 (account and checkout credentials), a locked OS credential store, feature_flag_disabled, and quote_required. Each heading is the literal code or message, so search finds it.
• Getting started (docs/getting-started.md): one path from install through login, a saved profile, a mutinynet wallet, a received 1,000 sats, and the same receive scripted with --json and jq.
• Changelog: CHANGELOG.md (Keep a Changelog style, hand-written) seeded with the changes since 0.1.0, including this stack.
• Keeping it current: a Changelog workflow fails a PR that changes src/cli.rs or api/commands.json without CHANGELOG.md, unless it has the no-changelog label. It reruns on label changes, so it is a separate workflow from CI. AGENTS.md states the rule and that the check is a heuristic: it fires on test-only src/cli.rs changes and misses user-visible changes in other files.
• PR template: asks for the changelog entry, the verification gate, the voltage-pr-review pass with tests for new error paths, the CRAP result (highest score and each function at or above 20), and each check not run with its reason, as AGENTS.md "Before you request review" requires.

Tests: troubleshooting must have an entry for every exit code and every CredentialProblem. The exit codes come from #28's error_kinds() and the problems from an exhaustive match, so a new kind or problem fails to compile until it has an entry. The docs example checks also read CHANGELOG.md and the tutorial, so the tutorial's commands parse and its changes pass as offline dry runs.

Changes after review

• The PR template asks for the review checks, the CRAP result, and checks not run.
• The exit-code test derives its codes from error_kinds() instead of a hard-coded list.
• Imports moved to module scope. tests/support/examples.rs is now rustfmt-clean; cargo fmt does not reach it because it is included with include!.
• New auth_url_mismatch entry (from feat(cli): report auth_url in context, time waits, and list items one per line #19). The 401/403 entries cover checkout tokens, and the quote_required entry matches feat(cli): name the next command in scope, credential, and login errors #24's wording.
• The no-changelog label now exists in the repository.

Not done here

• The tutorial is unverified against a live account. The plan's acceptance check is a new teammate following it on a test network.
• A staging guide waits for deployment URLs in profiles (plan 05).

Checks

On 574e085:

• cargo fmt --check, cargo clippy --locked --all-targets -- -D warnings, cargo machete --with-metadata: pass.
• cargo test --locked: 103 unit and 86 integration tests pass. VOLTAGE_BLESS=1 cargo test --locked leaves the tree clean.
• python3 scripts/check-coverage.py: pass.
• CRAP (scripts/crap-report.sh): no function at or above 20. Highest is api::wait_for_payment at 19.24, unchanged from or lower than master (19.66).
• Skill review: voltage-pr-review and the voltage-backend-engineering Rust checklist, against the stack base. Findings are fixed in this commit.

Not run: the staging and live tutorials, which need credentials.

[tee8z]

One line repeats the auth_url_mismatch advice that #19 fixed; everything else matched the code (exit codes, quoted errors, anchors), and the changelog workflow and PR template look right. CRAP on 7c99bae: no function at or above 20, highest api::wait_for_payment at 19.24; the full test suite passed in the same run.