Earlier verification records
Claude Code: Earlier review dates, version baselines, and findings, with a link to the current verification scope.
Key takeaways
- This page preserves earlier review records. Use the detail guides and complete list for current behavior. Earlier minimum-version notes do not establish a newly verified introduction version.
September 22, 2026 second audit
The audit covers 239 pages across both books and languages: full command declarations, official URL destinations and anchors, internal links, and MDX compilation. Redirected documentation links now use the current destinations.
The official table now has 114 entries including /update-config. Corrections cover the retired /agents wizard, /review alias, configuration/MCP/reload/effort arguments, the /design minimum, and ultrareview billing and whole-repository fallback.
Checked 131 real external URLs (124 without fragments), with no 404s among source links. The OpenAI announcement was readable through a separate web fetch. The npm website and Claude routines console returned 403 here and remain unverified, rather than being counted as valid or broken. Reference checks do not establish account-specific runtime availability.
Running verification
Local checks validate complete declarations, CLI tables, locale coverage, and navigation against the reviewed snapshot. Live comparison fails when official names, arguments, types, descriptions, or latest releases change; source failures also fail. Behavior and availability still require editorial review.
pnpm --filter handbook check:claude-command-master
pnpm --filter handbook test:command-sources
pnpm --filter handbook check:command-sourcesOfficial command audit — September 22, 2026
Reconciled 114 Commands entries, 34 CLI invocation forms, and 79 common CLI flags. The release baseline is v2.1.278 (September 19, 2026) in the official changelog. Older npm channel values below describe their recorded dates, not the current channels.
-
Restored
/output-style [style]as a current documented command, with a reference row and usage page. -
Added
/claude-api build-evalandhillclimb(v2.1.259+), and/import cursor(v2.1.265+). -
Added plugin install/update
--accept-command <sha256>(v2.1.271),/plugin install <plugin> --marketplace <source>(v2.1.275), and runner--drain-marker-file <path>(v2.1.271). -
Updated
/configProject instructions andAGENTS.mdfallback (v2.1.277),/statusAuto mode server (v2.1.278), and/fastin Remote sessions (v2.1.271). -
Added
/update-config [request], corrected obsolete/agentsand/reviewguidance, the/designminimum (v2.1.265), and CLI stream/budget details.
All slash commands · Terminal CLI reference
Sources: Commands, CLI reference, Changelog.
The sections below retain earlier update records.
Cross-checked all 112 entries in the official Commands table. Added /design and /skill-doctor to both locales and the checker. /output-style remains release-tracked.
September 12, 2026 release refresh
Updated for Claude Code v2.1.263–v2.1.269 (September 6–11). npm checked on September 12, 2026 KST: latest/next=2.1.269, stable=2.1.236. Keep latest-feature evaluation separate from a team's stable-channel rollout.
- Plugins:
claude plugin evalproduces JSON and HTML evaluation reports; install, uninstall, update, enable, and disable accept--json.--plugin-dircan point to a parent directory containing multiple plugins. - Commands and settings:
/output-style [name]works over Remote Control and in cloud/headless sessions.maxEffortLevelcaps effort globally or per model undermodelSettings.bashEditDiffEnabledadds file-edit diffs to Bash results. - Recovery and compatibility: 2.1.266 fixes forced gateway sign-in; 2.1.268 fixes HTTP 400s on Anthropic-compatible endpoints introduced in 2.1.265.
/goalretries with backoff or pauses with a reason after API, network, or usage-limit failures. - Permissions and operations: fixes cover symlink deny/ask rules, Bash
teewrite boundaries, and secrets in MCP/plugin errors. WebFetch has a default 300-second deadline./ultrareview --postposts the PR comment directly when findings arrive.
Sources: official changelog, npm dist-tags.
This refresh compares official release text with npm channels. It does not claim local CLI execution or live feature testing. Tables below retain prior dated checks; the channel values above supersede them.
Source Baseline
Checked on 2026-09-05 against:
-
Claude Fable 5.1 overview: https://platform.claude.com/docs/en/models/fable-5-1/overview
-
What's new in Claude Fable 5.1: https://platform.claude.com/docs/en/models/fable-5-1/whats-new-fable-5-1
-
Migrating to Claude Fable 5.1: https://platform.claude.com/docs/en/models/fable-5-1/migration-guide
-
Fable 5.1 and Mythos 5.1 announcement: https://www.anthropic.com/claude-fable-and-mythos-5-1
-
Official Commands reference: https://code.claude.com/docs/en/commands
-
Official CLI reference: https://code.claude.com/docs/en/cli-reference
-
Official Changelog: https://code.claude.com/docs/en/changelog
-
Official Model configuration: https://code.claude.com/docs/en/model-config
-
Official Agent View guide: https://code.claude.com/docs/en/agent-view
-
Official Skills guide: https://code.claude.com/docs/en/skills
-
Official Fast mode guide: https://code.claude.com/docs/en/fast-mode
-
Official Routines guide: https://code.claude.com/docs/en/routines
-
Claude Fable 5 and Claude Mythos 5: https://platform.claude.com/docs/en/models/fable-5/introducing-claude-fable-5-and-claude-mythos-5
-
Fable refusals and fallback: https://platform.claude.com/docs/en/build-with-claude/refusals-and-fallback
-
Fable fallback credit: https://platform.claude.com/docs/en/build-with-claude/fallback-credit
-
Prompting Claude Fable 5: https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-fable-5
-
Usage credits: https://support.claude.com/en/articles/12429409-manage-usage-credits-for-paid-claude-plans
-
Claude Sonnet 5 announcement: https://www.anthropic.com/news/claude-sonnet-5
-
Claude models overview: https://platform.claude.com/docs/en/models/overview
-
Claude Opus 4.8 announcement: https://www.anthropic.com/news/claude-opus-4-8
-
npm package: https://www.npmjs.com/package/@anthropic-ai/claude-code
Locale Checks
/en/books/claude-code-command-masterrenders English content only.- The English root meta registers
claude-code-command-master. - Home preview chapters are loaded with
locale=en. - Unsupported English books remain translation requests instead of falling back to Korean content.
- The locale switcher should allow English for this book once the page exists in both locales.
Editorial Checks
| Area | Status |
|---|---|
| Cover metadata | Complete |
| Command catalog | Complete |
| Core command grouping | Complete |
| Skills/custom commands | Complete |
| Web automation commands | Complete |
| Operations and safety commands | Complete |
| Official Commands inventory | Complete: all 110 rows in both locale reference pages |
| Update log | Complete |
Current Baseline Checks
| Area | Baseline |
|---|---|
| Claude Code version | v2.1.261 latest and next; v2.1.236 stable |
| v2.1.252–261 changes | Fable 5.1 as default Fable (v2.1.257), gateway aliases kept on Fable 5, /effort session-only s, CLAUDE_CODE_SUBAGENT_MODEL_FORCE, the v2.1.260 fullscreen /diff panel, Fable 5.1 picker/prompt-cache/[1m] fixes, and 1M auto-compact timing. /skill-doctor, managedMcpServers, and --permission-prompts none are changelog-only so far and stay out of the 110-row reference |
| Fable 5.1 | Released 2026-09-01 as claude-fable-5-1; Claude Code v2.1.255+; same pricing as Fable 5 with $0.25/MTok cache reads; not default on any plan; Fable 5 via /model claude-fable-5 |
| Fable 5.1 API deltas | Forced tool_choice returns 400; thinking blocks bound to Fable 5.1; editing earlier turns invalidates thinking (enforced for accounts created on/after 2026-08-31); per-message effort, clear_at, display: "updates", and provenance are additive |
| Classifier fallback table | Fable 5.1/5 biology to Opus 5, cybersecurity to Opus 4.8; Opus 5 cybersecurity to Opus 4.8, biology refused; switchModelsOnFlag: false; third-party needs ANTHROPIC_DEFAULT_OPUS_MODEL |
| Mythos 5.1 | Same specs and pricing as Fable 5.1 with a different safeguard level; Project Glasswing approved customers only; not selectable via /model |
| Fast mode | Opus 5 and Opus 4.8 only; Fable 5.1 and Fable 5 unsupported; Opus 4.7 fast mode removed 2026-07-24 |
| v2.1.248–251 changes | Model-switch hooks, usage and prompt-cache telemetry, restricted mode, background-session CLI, workflow authoring, symlink/path hardening, and agent/Remote Control reliability |
| v2.1.239–247 changes | Per-loop usage, Auto-mode rules, Claude API upgrade/cost optimization, feedback drafts, partial subagent continuation, headless recovery, and security hardening |
| v2.1.221–237 changes | Interactive fork/background spawn, cross-session messaging, self-hosted runner, GitLab MR, default model, Concise output, security and reliability fixes |
| Default model env | ANTHROPIC_DEFAULT_MODEL is a low-precedence new-session default from v2.1.236 |
| Task/Todo tools | Disabled by default on newer models from v2.1.233; compatibility flag documented |
| v2.1.206–220 changes | Opus 5; /fork//subtask split; manual-only review/verify; forked-skill background default; agent limits; sandbox, hooks, and headless MCP-error updates |
| v2.1.202–205 changes | Workflow sizing/OTel, fast /review split, Agent View/background/MCP roots, headless hook streaming, full /doctor checkup, Auto-mode and JSON Schema fixes |
| v2.1.198–201 changes | Subagents background-by-default (v2.1.198), /dataviz skill, Claude in Chrome GA, default permission mode renamed "Manual" (v2.1.200), Sonnet 5 mid-conversation system role removed (v2.1.201) |
| Fable path (history) | Fable 5 on Claude Code v2.1.170+ (released June 9, 2026); fable resolved to Fable 5 before v2.1.255 and to Fable 5.1 afterwards; never the default |
| Fable 5 redeployment | Suspended 2026-06-12 under US export controls; restored globally 2026-07-01 with an improved safety classifier. Mythos 5 (no classifiers) is Project Glasswing invitation-only and not selectable via /model |
| Fable plan access | Temporary subscription allowance ended 2026-07-07; current Pro/Max/Team and seat-based Enterprise access requires usage credits, while API consumption is billed separately |
| Fable API refusal | HTTP 200 with stop_reason: "refusal"; monitor it separately from transport and HTTP errors |
| Fable API fallback | Server-side fallbacks, SDK middleware, and manual retry documented separately from Claude Code availability fallback |
| Fallback billing | Pre-output refusals are not billed; fallback credit avoids duplicate prompt-cache write cost on a model switch |
| Fable prompting | Avoid internal-reasoning extraction instructions; use concise rationale and structured adaptive-thinking blocks |
| Sonnet path | Sonnet 5 on Claude Code v2.1.197+; default for Pro, Team Standard, and Enterprise subscription seats |
| Latest Opus path | Opus 5 with native 1M context on Claude Code v2.1.219+ |
| Anthropic API aliases | fable → Fable 5.1 (v2.1.255+), best → latest Fable when available otherwise latest Opus, opus → Opus 5, sonnet → Sonnet 5; Claude apps gateway sessions keep fable/best on Fable 5 |
| Recent command deltas | /advisor accepts fable; /cd relocates the session and storage; --safe-mode isolates customization while keeping auth/model/permissions |
| Model fallback | Fable classifier fallback to Opus; fallbackModel chains for availability failures |
| Skill controls | --disable-slash-commands; disableBundledSkills; CLAUDE_CODE_DISABLE_BUNDLED_SKILLS |
| CI reliability | Self-hosted runner post-session hook; v2.1.170 transcript persistence fix for VS Code integrated terminals and inherited Claude Code env vars |
| Background work | Agent View, /background//bg, claude --bg, claude --bg --exec |
| Dynamic work | /goal and /workflows |
| Usage continuation | /usage-credits; /extra-usage treated as legacy |
| Review split | /code-review for correctness, /simplify for cleanup |
| MCP operations | claude mcp login/logout; CLAUDE_CODE_MCP_TOOL_IDLE_TIMEOUT |
| Sandbox operations | sandbox.credentials, sandbox.network.strictAllowlist, sandbox.filesystem.disabled, destructive-command and Windows fixes |
| Agent limits | Three-level nesting and 20 concurrent by default; old 200-per-session total removed in v2.1.224 |
| Model warnings | Main model restriction notes and stronger deprecated-model warnings |
Coverage Model
The English edition groups detailed guidance by workflow instead of translating every Korean detail page one-to-one. Coverage is still exhaustive at the inventory level: the complete command reference contains all 114 official Commands rows checked on 2026-09-22, including the removed entries retained in the source table.