This page is generated from the same command registry that prints accessibility-audit --help. Regenerate it with node scripts/build-cli-reference.mjs --write after changing a command.
The Japanese blocks are the actual --locale ja output. Text without a Japanese translation stays in English, as it does in the CLI.
このページは、accessibility-audit --helpと同じコマンド定義から生成しています。コマンドを変更したらnode scripts/build-cli-reference.mjs --writeで再生成してください。
日本語版は--locale jaの実際の出力です。日本語訳がない説明は、CLIと同じく英語のまま表示します。
| Command / コマンド | Summary | 概要 |
|---|---|---|
artifact |
Wrap a completed payload in a run-bound candidate, or check it before registration. | 完成したpayloadをrunに紐づく候補artifactへ包むか、登録前に検査します。 |
review-queue |
Create a target-bound human review candidate from registered observations and explicit requirements. | 登録済みobservationと明示した条項から、対象に紐づく人手レビュー候補を作成します。 |
audit-bundle |
Prepare or verify offline commitments to audit files and a signed predecessor chain. | 保存した監査ファイルと前段の署名を、オフラインで準備・照合します。 |
human-review |
Export/import human review worksheets; prepare, verify or apply portable review records. | 人手レビューのworksheetを書き出し・取り込み、持ち運べるレビュー記録を準備・照合・適用します。 |
network-policy |
Propose concrete target and official-source network scopes without granting or using network access. | network接続を許可も使用もせずに、対象と公式資料のnetwork範囲を具体的に提案します。 |
import |
Import native axe-core results as private, run-bound screening candidates. | axe-coreの結果を、runに紐づく非公開のscreening候補として取り込みます。 |
bind-targets |
Remeasure and bind a target inventory once, before any artifact registration. | artifactを登録する前に一度だけ、対象inventoryを再測定して固定します。 |
compare-targets |
Compare the measured identities of two bound runs without contacting live targets. | 実際の対象へ接続せずに、固定済みの2つのrunの測定済み識別情報を比較します。 |
capture-targets |
Measure file, Git, HTTP or saved web-state identities into a private run companion. | file、Git、HTTP、保存済みWeb状態の識別情報を測定し、非公開のrun付属ファイルへ記録します。 |
bind-evidence |
Bind an existing private evidence file to one draft screening observation. | 既存の非公開証拠fileを、下書きのscreening observation 1件へ紐づけます。 |
compare-runs |
Compare a linked retest with its predecessor using registered findings, human outcomes and screening signals. | 登録済みfinding、人手outcome、screening signalを使い、再検査runを前のrunと比較します。 |
compare-evidence |
Compare validated saved evidence before and after a change, for private review. | 変更前後の検証済み保存証拠を、非公開レビュー用に比較します。 |
lifecycle |
Create immutable private finding-management revisions linked to registered audit findings. | 登録済みの監査findingに紐づく、不変で非公開のfinding管理revisionを作成します。 |
status |
Read run state, evidence coverage, validation, successor warnings and next operations. | 監査の状態、証拠の範囲、検証結果、後続runと次の操作を読み取ります。 |
init |
Create a new immutable audit run. | 新しい不変のaudit runを作成します。 |
assessment |
Create a validator-valid assessment record or an explicit placeholder template. | validator-validなassessment recordまたは明示的なplaceholder templateを作成します。 |
scan-web |
Run rule-based browser checks and create a compact AI review context. | rule-based browser checkを実行し、圧縮したAI review contextを作成します。 |
profiles |
List active standards profiles, counts, groups, sources, and claim ceilings. | activeな規格profile、件数、group、source、claim上限を一覧表示します。 |
requirements |
List, search, or show WCAG/JIS requirements without requiring internal IDs. | 内部IDを知らなくてもWCAG・JIS条項を一覧・検索・表示できます。 |
requirement |
Show one registered requirement by exact internal ID (legacy-compatible entry point). | 正確な内部IDで登録済み条項を1件表示します（legacy互換入口）。 |
preflight-web |
Measure Web inspection capabilities on an isolated fixture without inspecting a target. | 対象を検査せずに、分離したfixtureでWeb検査の実行能力を測定します。 |
doctor |
Diagnose Node, package, registry, distribution, and optional browser capabilities. | Node、package、registry、distribution、任意browser capabilityを診断します。 |
non-web-review |
Create, validate, report, and compare private participation reviews for non-Web targets. | Web以外の対象について、非公開の参加レビューを作成・検証・報告・比較します。 |
screen-reader-checklist |
Show supporting checks for stateful UI and screen-reader behavior. | 状態を持つUIとscreen-reader動作の補助checkを表示します。 |
validate-run |
Validate an immutable audit run and write a new validation record. | 不変のaudit runを検証し、新しいvalidation recordを書き出します。 |
validate-assessment |
Validate an assessment and print its coverage and claim guard result. | assessmentを検証し、coverageとclaim guard結果を表示します。 |
register |
Register one validated artifact in a new audit-run version. | 検証済みartifactを新しいaudit-run版へ1件登録します。 |
merge |
Merge registered artifacts into a new assessment. | 登録済みartifactを新しいassessmentへ統合します。 |
report |
Render a profile-aware, provenance-explicit Markdown report from a validated standalone or run-backed assessment. | 検証済みstandaloneまたはrun-backed assessmentから、profile対応・provenance明示のレポートを生成します。 |
retest |
Create a fresh audit run from a registered authorized or externally declared change. | 認可済み変更が完了したpredecessorから新しいaudit runを作成します。 |
Information Accessibility Audit CLI
Usage:
accessibility-audit <command> [options]
accessibility-audit --version
Global options:
--help, -h Show this help.
--version, -V Show package and installed contract versions.
--locale <ja|en> Human-readable locale: ja or en.
Commands:
artifact Wrap a completed payload in a run-bound candidate, or check it before registration.
review-queue Create a target-bound human review candidate from registered observations and explicit requirements.
audit-bundle Prepare or verify offline commitments to audit files and a signed predecessor chain.
human-review Export/import human review worksheets; prepare, verify or apply portable review records.
network-policy Propose concrete target and official-source network scopes without granting or using network access.
import Import native axe-core results as private, run-bound screening candidates.
bind-targets Remeasure and bind a target inventory once, before any artifact registration.
compare-targets Compare the measured identities of two bound runs without contacting live targets.
capture-targets Measure file, Git, HTTP or saved web-state identities into a private run companion.
bind-evidence Bind an existing private evidence file to one draft screening observation.
compare-runs Compare a linked retest with its predecessor using registered findings, human outcomes and screening signals.
compare-evidence Compare validated saved evidence before and after a change, for private review.
lifecycle Create immutable private finding-management revisions linked to registered audit findings.
status Read run state, evidence coverage, validation, successor warnings and next operations.
init Create a new immutable audit run.
assessment Create a validator-valid assessment record or an explicit placeholder template.
scan-web Run rule-based browser checks and create a compact AI review context.
profiles List active standards profiles, counts, groups, sources, and claim ceilings.
requirements List, search, or show WCAG/JIS requirements without requiring internal IDs.
requirement Show one registered requirement by exact internal ID (legacy-compatible entry point).
preflight-web Measure Web inspection capabilities on an isolated fixture without inspecting a target.
doctor Diagnose Node, package, registry, distribution, and optional browser capabilities.
non-web-review Create, validate, report, and compare private participation reviews for non-Web targets.
screen-reader-checklist Show supporting checks for stateful UI and screen-reader behavior.
validate-run Validate an immutable audit run and write a new validation record.
validate-assessment Validate an assessment and print its coverage and claim guard result.
register Register one validated artifact in a new audit-run version.
merge Merge registered artifacts into a new assessment.
report Render a profile-aware, provenance-explicit Markdown report from a validated standalone or run-backed assessment.
retest Create a fresh audit run from a registered authorized or externally declared change.
This command is a thin, read-only control-plane wrapper around the installed skill runtime.
It does not evaluate conformance by itself and does not expose target mutation.
Run accessibility-audit <command> --help for command-specific usage.
Recommended private output: audit-runs/<run-id>/; explicitly export and review public files before delivery.
情報アクセシビリティ監査CLI
使用方法:
accessibility-audit <command> [options]
accessibility-audit --version
共通オプション:
--help, -h このヘルプを表示します。
--version, -V パッケージとインストール済み契約の版を表示します。
--locale <ja|en> 表示言語。jaまたはenを指定します。
コマンド:
artifact 完成したpayloadをrunに紐づく候補artifactへ包むか、登録前に検査します。
review-queue 登録済みobservationと明示した条項から、対象に紐づく人手レビュー候補を作成します。
audit-bundle 保存した監査ファイルと前段の署名を、オフラインで準備・照合します。
human-review 人手レビューのworksheetを書き出し・取り込み、持ち運べるレビュー記録を準備・照合・適用します。
network-policy network接続を許可も使用もせずに、対象と公式資料のnetwork範囲を具体的に提案します。
import axe-coreの結果を、runに紐づく非公開のscreening候補として取り込みます。
bind-targets artifactを登録する前に一度だけ、対象inventoryを再測定して固定します。
compare-targets 実際の対象へ接続せずに、固定済みの2つのrunの測定済み識別情報を比較します。
capture-targets file、Git、HTTP、保存済みWeb状態の識別情報を測定し、非公開のrun付属ファイルへ記録します。
bind-evidence 既存の非公開証拠fileを、下書きのscreening observation 1件へ紐づけます。
compare-runs 登録済みfinding、人手outcome、screening signalを使い、再検査runを前のrunと比較します。
compare-evidence 変更前後の検証済み保存証拠を、非公開レビュー用に比較します。
lifecycle 登録済みの監査findingに紐づく、不変で非公開のfinding管理revisionを作成します。
status 監査の状態、証拠の範囲、検証結果、後続runと次の操作を読み取ります。
init 新しい不変のaudit runを作成します。
assessment validator-validなassessment recordまたは明示的なplaceholder templateを作成します。
scan-web rule-based browser checkを実行し、圧縮したAI review contextを作成します。
profiles activeな規格profile、件数、group、source、claim上限を一覧表示します。
requirements 内部IDを知らなくてもWCAG・JIS条項を一覧・検索・表示できます。
requirement 正確な内部IDで登録済み条項を1件表示します（legacy互換入口）。
preflight-web 対象を検査せずに、分離したfixtureでWeb検査の実行能力を測定します。
doctor Node、package、registry、distribution、任意browser capabilityを診断します。
non-web-review Web以外の対象について、非公開の参加レビューを作成・検証・報告・比較します。
screen-reader-checklist 状態を持つUIとscreen-reader動作の補助checkを表示します。
validate-run 不変のaudit runを検証し、新しいvalidation recordを書き出します。
validate-assessment assessmentを検証し、coverageとclaim guard結果を表示します。
register 検証済みartifactを新しいaudit-run版へ1件登録します。
merge 登録済みartifactを新しいassessmentへ統合します。
report 検証済みstandaloneまたはrun-backed assessmentから、profile対応・provenance明示のレポートを生成します。
retest 認可済み変更が完了したpredecessorから新しいaudit runを作成します。
このコマンドは、インストール済みskill runtimeを操作する薄い読み取り専用control-plane wrapperです。
このコマンドだけで適合性を判定せず、監査対象を変更する機能も公開しません。
各コマンドの詳しい使い方は accessibility-audit <command> --help で確認できます。
内部成果物の推奨保存先: audit-runs/<run-id>/（公開用ファイルは確認後に明示的に書き出してください）。
Wrap a completed payload in a run-bound candidate, or check it before registration.
Usage:
accessibility-audit artifact init --run <run.json> --type <type> --payload <payload.json> [--input <registered-ART-id>] [--artifact-id <ART-id>] --output <artifacts/new-candidate.json>
accessibility-audit artifact validate --run <run.json> --artifact <artifacts/candidate.json>
Notes:
- Types: screening-observations, human-review-queue, declared-human-review, remediation-plan, audit-context, participant-usability-observation, declared-change-record. --input is repeatable.
- For audit-context, also provide --role <declared_context_reviewer|declared_context_owner> --evidence-file <saved-file> --target-ref <declared-target> --captured-at <RFC3339>. See references/audit-context.md.
- For participant-usability-observation, use --role declared_participant_facilitator with the same saved-evidence options. See references/participant-usability-observation.md.
- For declared-change-record, use --input <remediation-plan-id>, --role declared_change_reviewer|declared_change_owner, --after-inventory <captured.json>, and the same saved-evidence options. See references/declared-change-record.md.
- The CLI fills envelope metadata and omitted payload schema_version. Supply actual observations, review declarations or proposals in --payload; none are invented.
- Human review payloads must be the reviewer's actual declaration; AI agents must not write human outcomes or evidence on their behalf. Metadata does not authenticate identity.
- Creates a private candidate only; register performs live target checks. Prefer review-queue and human-review export/import for guided authoring. See references/agent-orchestration.md.
完成したpayloadをrunに紐づく候補artifactへ包むか、登録前に検査します。
使用方法:
accessibility-audit artifact init --run <run.json> --type <type> --payload <payload.json> [--input <registered-ART-id>] [--artifact-id <ART-id>] --output <artifacts/new-candidate.json>
accessibility-audit artifact validate --run <run.json> --artifact <artifacts/candidate.json>
注意:
- Types: screening-observations, human-review-queue, declared-human-review, remediation-plan, audit-context, participant-usability-observation, declared-change-record. --input is repeatable.
- For audit-context, also provide --role <declared_context_reviewer|declared_context_owner> --evidence-file <saved-file> --target-ref <declared-target> --captured-at <RFC3339>. See references/audit-context.md.
- For participant-usability-observation, use --role declared_participant_facilitator with the same saved-evidence options. See references/participant-usability-observation.md.
- For declared-change-record, use --input <remediation-plan-id>, --role declared_change_reviewer|declared_change_owner, --after-inventory <captured.json>, and the same saved-evidence options. See references/declared-change-record.md.
- The CLI fills envelope metadata and omitted payload schema_version. Supply actual observations, review declarations or proposals in --payload; none are invented.
- Human review payloads must be the reviewer's actual declaration; AI agents must not write human outcomes or evidence on their behalf. Metadata does not authenticate identity.
- Creates a private candidate only; register performs live target checks. Prefer review-queue and human-review export/import for guided authoring. See references/agent-orchestration.md.
Create a target-bound human review candidate from registered observations and explicit requirements.
Usage:
accessibility-audit review-queue --run <run.json> --artifact-id <ART-id> --output <artifacts/new-queue.json> [--scope screening|profile_all] [--requirement <profile-requirement-id>]
Notes:
- --requirement is repeatable. Priorities start unprioritized; review reasons, target locations and required state before registration.
- Creates a private candidate only. It does not perform human review, register evidence, or change the audit run. See references/human-review-queue.md.
登録済みobservationと明示した条項から、対象に紐づく人手レビュー候補を作成します。
使用方法:
accessibility-audit review-queue --run <run.json> --artifact-id <ART-id> --output <artifacts/new-queue.json> [--scope screening|profile_all] [--requirement <profile-requirement-id>]
注意:
- --requirement is repeatable. Priorities start unprioritized; review reasons, target locations and required state before registration.
- Creates a private candidate only. It does not perform human review, register evidence, or change the audit run. See references/human-review-queue.md.
Prepare or verify offline commitments to audit files and a signed predecessor chain.
Usage:
accessibility-audit audit-bundle prepare --root <directory> --run <run.json> --assessment <assessment.json> --report <report.md|html> --output <new-private-record.json>
accessibility-audit audit-bundle verify --root <directory> --record <private-record.json> [trust options]
Options:
--report <file> Prepare: 1–16 explicit reports; repeat for multiple formats.
--attachment <file> Prepare: explicitly retained source/capture file; repeatable.
--predecessor <record.json> Repeat for earlier signed records; prepare takes the immediate predecessor first.
--require-complete-chain Reject any superseded run without a predecessor signature.
--trust-policy <policy.json> Recipient-selected external signer policy.
--trust-policy-sha256 <SHA-256> Independently trusted canonical policy digest; required with policy.
--minimum-assurance <level> Verify: unsigned, self_signed, signed, organization_attested or independent; applies to every link.
--expected-subject-sha256 <SHA-256> Verify: recipient-selected current bundle digest, to reject an older valid record.
Notes:
- All committed files must be inside --root. Paths supplied on the CLI resolve from the working directory.
- No keys are created, no data is uploaded, and no conformance tier is elevated. Historical file bytes and trusted timestamps are not verified.
- See references/audit-bundle-attestation.md for the signed format, offline trust and archival limits.
保存した監査ファイルと前段の署名を、オフラインで準備・照合します。
使用方法:
accessibility-audit audit-bundle prepare --root <directory> --run <run.json> --assessment <assessment.json> --report <report.md|html> --output <new-private-record.json>
accessibility-audit audit-bundle verify --root <directory> --record <private-record.json> [trust options]
オプション:
--report <file> 準備時に署名対象へ含めるレポート。1〜16件を個別に指定します。
--attachment <file> 準備時に含める保存済みソースや取得ファイル。複数指定できます。
--predecessor <record.json> 過去の署名済み記録。準備時は直前の記録を最初に指定します。
--require-complete-chain 前段の署名がないsupersedes履歴を拒否します。
--trust-policy <policy.json> 受領者が独立に選んだ署名者の信頼方針。
--trust-policy-sha256 <SHA-256> 別途信頼した方針の正規化SHA-256。方針と対で必須です。
--minimum-assurance <level> 検証時の最低水準。unsigned、self_signed、signed、organization_attested、independent。全段へ適用します。
--expected-subject-sha256 <SHA-256> 受領者が別途確認した今回の署名対象ハッシュ。古い有効記録への置換を拒否します。
注意:
- 署名対象ファイルはすべて--root内に置きます。CLIの相対パスは作業ディレクトリ基準です。
- 鍵の生成、送信、適合主張の昇格は行いません。過去の実ファイルと信頼できる時刻は未確認です。
- 形式、外部の信頼方針、保存上の制約はreferences/audit-bundle-attestation.mdを参照してください。
Export/import human review worksheets; prepare, verify or apply portable review records.
Usage:
accessibility-audit human-review prepare --assessment <assessment.json> --review <declared-review.json> --reviewer-id <id> --output <new-record.json>
accessibility-audit human-review export --run <run.json> --queue <registered-queue-id> --format xlsx|csv|markdown --output <new-sheet>
accessibility-audit human-review import --run <run.json> --queue <registered-queue-id> --input <completed-sheet> --output <new-artifact.json> [--artifact-id <id>]
accessibility-audit human-review prepare --run <run.json> --artifact-id <id> --reviewer-id <id> --output <new-record.json>
accessibility-audit human-review verify --record <record.json> --assessment <assessment.json> [trust options]
accessibility-audit human-review apply --record <record.json> --assessment <assessment.json> --output <new-assessment.json> [trust options]
Notes:
- Trust options: --trust-policy <recipient-selected.json> --trust-policy-sha256 <independent canonical SHA-256> [--minimum-assurance signed].
- Verify also accepts --run and --artifact-id. Run-backed results are applied by merge --review-record.
- No signing keys are created. Stored flags, self-signed keys and declarations never authenticate a person. See references/reviewer-assurance.md.
人手レビューのworksheetを書き出し・取り込み、持ち運べるレビュー記録を準備・照合・適用します。
使用方法:
accessibility-audit human-review prepare --assessment <assessment.json> --review <declared-review.json> --reviewer-id <id> --output <new-record.json>
accessibility-audit human-review export --run <run.json> --queue <registered-queue-id> --format xlsx|csv|markdown --output <new-sheet>
accessibility-audit human-review import --run <run.json> --queue <registered-queue-id> --input <completed-sheet> --output <new-artifact.json> [--artifact-id <id>]
accessibility-audit human-review prepare --run <run.json> --artifact-id <id> --reviewer-id <id> --output <new-record.json>
accessibility-audit human-review verify --record <record.json> --assessment <assessment.json> [trust options]
accessibility-audit human-review apply --record <record.json> --assessment <assessment.json> --output <new-assessment.json> [trust options]
注意:
- Trust options: --trust-policy <recipient-selected.json> --trust-policy-sha256 <independent canonical SHA-256> [--minimum-assurance signed].
- Verify also accepts --run and --artifact-id. Run-backed results are applied by merge --review-record.
- No signing keys are created. Stored flags, self-signed keys and declarations never authenticate a person. See references/reviewer-assurance.md.
Propose concrete target and official-source network scopes without granting or using network access.
Usage:
accessibility-audit network-policy [--target <URL>] [--exact-target <URL>] [--include-official-sources true] [--profile <id>] [--method GET|HEAD] [--allow-localhost true] [--output <new-policy.json>]
Notes:
- Targets, exact targets and methods are repeatable. Review the proposal before passing its file to init --network-policy.
- Official sources are exact registered URLs in a separate purpose scope. A policy file never grants caller authority.
network接続を許可も使用もせずに、対象と公式資料のnetwork範囲を具体的に提案します。
使用方法:
accessibility-audit network-policy [--target <URL>] [--exact-target <URL>] [--include-official-sources true] [--profile <id>] [--method GET|HEAD] [--allow-localhost true] [--output <new-policy.json>]
注意:
- Targets, exact targets and methods are repeatable. Review the proposal before passing its file to init --network-policy.
- Official sources are exact registered URLs in a separate purpose scope. A policy file never grants caller authority.
Import native axe-core results as private, run-bound screening candidates.
Usage:
accessibility-audit import axe --run <bound-run.json> --input <artifacts/axe.json> --output <artifacts/new-screening.json> [--target-ref <declared-URL>] [--configuration <artifacts/config.json>] [--record-output <artifacts/new-import.json>] [--artifact-id <ART-id>]
Notes:
- Preserves every result category, unknown rules and raw evidence. No profile pass/fail is inferred.
- Requires a bound saved web_state. Native JSON has caller-declared capture association; scan-web exports additionally bind the saved bundle hash.
- Creates candidates only. Register the artifact, then provide the required human-review queue before merge.
axe-coreの結果を、runに紐づく非公開のscreening候補として取り込みます。
使用方法:
accessibility-audit import axe --run <bound-run.json> --input <artifacts/axe.json> --output <artifacts/new-screening.json> [--target-ref <declared-URL>] [--configuration <artifacts/config.json>] [--record-output <artifacts/new-import.json>] [--artifact-id <ART-id>]
注意:
- Preserves every result category, unknown rules and raw evidence. No profile pass/fail is inferred.
- Requires a bound saved web_state. Native JSON has caller-declared capture association; scan-web exports additionally bind the saved bundle hash.
- Creates candidates only. Register the artifact, then provide the required human-review queue before merge.
Remeasure and bind a target inventory once, before any artifact registration.
Usage:
accessibility-audit bind-targets --run <unbound-run.json> --targets <artifacts/targets.json> [--allow-origin <origin>] [--allow-url <exact-URL>] [--allow-localhost true] [--network-log-output <artifacts/network.json>] --output <new-bound-run.json>
Notes:
- The output stays beside the input run. Changing a bound inventory requires a fresh run. HTTP origins must be explicitly authorized again.
artifactを登録する前に一度だけ、対象inventoryを再測定して固定します。
使用方法:
accessibility-audit bind-targets --run <unbound-run.json> --targets <artifacts/targets.json> [--allow-origin <origin>] [--allow-url <exact-URL>] [--allow-localhost true] [--network-log-output <artifacts/network.json>] --output <new-bound-run.json>
注意:
- The output stays beside the input run. Changing a bound inventory requires a fresh run. HTTP origins must be explicitly authorized again.
Compare the measured identities of two bound runs without contacting live targets.
Usage:
accessibility-audit compare-targets --before <run.json> --after <run.json> --output <after-artifacts/new-target-comparison.json>
Notes:
- Private output describes identity and environment changes, not accessibility outcomes. Both runs require measured target inventories.
実際の対象へ接続せずに、固定済みの2つのrunの測定済み識別情報を比較します。
使用方法:
accessibility-audit compare-targets --before <run.json> --after <run.json> --output <after-artifacts/new-target-comparison.json>
注意:
- Private output describes identity and environment changes, not accessibility outcomes. Both runs require measured target inventories.
Measure file, Git, HTTP or saved web-state identities into a private run companion.
Usage:
accessibility-audit capture-targets --run <run.json> --specs <target-specs.json> [--after-version <new-version>] [--allow-origin <origin>] [--allow-url <exact-URL>] [--allow-localhost true] [--network-log-output <artifacts/network.json>] --output <artifacts/new-targets.json>
Notes:
- Target specifications must cover exactly the declared run references. Local relative paths resolve beside the run manifest.
- HTTP requires a concrete run policy, explicit caller origin/URL authorization and --network-log-output. Every redirect is checked. No credentials, cookies or JavaScript are used.
- This creates an unregistered private companion. Use bind-targets before registering observations; --after-version instead measures a declared external change. See references/measured-targets.md.
file、Git、HTTP、保存済みWeb状態の識別情報を測定し、非公開のrun付属ファイルへ記録します。
使用方法:
accessibility-audit capture-targets --run <run.json> --specs <target-specs.json> [--after-version <new-version>] [--allow-origin <origin>] [--allow-url <exact-URL>] [--allow-localhost true] [--network-log-output <artifacts/network.json>] --output <artifacts/new-targets.json>
注意:
- Target specifications must cover exactly the declared run references. Local relative paths resolve beside the run manifest.
- HTTP requires a concrete run policy, explicit caller origin/URL authorization and --network-log-output. Every redirect is checked. No credentials, cookies or JavaScript are used.
- This creates an unregistered private companion. Use bind-targets before registering observations; --after-version instead measures a declared external change. See references/measured-targets.md.
Bind an existing private evidence file to one draft screening observation.
Usage:
accessibility-audit bind-evidence --run <run.json> --artifact <draft.json> --observation <SCREEN-id> --file <saved-file> --type <dom_snapshot|accessibility_tree|screenshot|interaction_log|network_log|other> --target-ref <declared-target> --captured-at <RFC3339> [--snapshot-id <id>] --output <new-artifact.json>
Notes:
- Drafts use screening-observations 4.0.0 and evidence_refs arrays. Bind every E1 observation before registration.
- Files and outputs stay within the artifact root. The command preserves observations and evidence levels; it does not inspect targets or authenticate the capture.
既存の非公開証拠fileを、下書きのscreening observation 1件へ紐づけます。
使用方法:
accessibility-audit bind-evidence --run <run.json> --artifact <draft.json> --observation <SCREEN-id> --file <saved-file> --type <dom_snapshot|accessibility_tree|screenshot|interaction_log|network_log|other> --target-ref <declared-target> --captured-at <RFC3339> [--snapshot-id <id>] --output <new-artifact.json>
注意:
- Drafts use screening-observations 4.0.0 and evidence_refs arrays. Bind every E1 observation before registration.
- Files and outputs stay within the artifact root. The command preserves observations and evidence levels; it does not inspect targets or authenticate the capture.
Compare a linked retest with its predecessor using registered findings, human outcomes and screening signals.
Usage:
accessibility-audit compare-runs --before <old-run.json> --after <retest-run.json> --output <after-artifacts/delta.json> --report <after-artifacts/delta.md> [--mapping <private-mapping.json>]
Notes:
- Both outputs stay in the successor private artifact root. Optional mapping contains finding_ids and screening_ids objects from old ID to new ID.
- A missing or inconclusive successor review is not resolved. Criterion passes do not prove that every old location was repaired; screening is separate from profile outcomes.
登録済みfinding、人手outcome、screening signalを使い、再検査runを前のrunと比較します。
使用方法:
accessibility-audit compare-runs --before <old-run.json> --after <retest-run.json> --output <after-artifacts/delta.json> --report <after-artifacts/delta.md> [--mapping <private-mapping.json>]
注意:
- Both outputs stay in the successor private artifact root. Optional mapping contains finding_ids and screening_ids objects from old ID to new ID.
- A missing or inconclusive successor review is not resolved. Criterion passes do not prove that every old location was repaired; screening is separate from profile outcomes.
Compare validated saved evidence before and after a change, for private review.
Usage:
accessibility-audit compare-evidence --before <run.json> --after <run.json> --output <after-artifacts/new-private-comparison.json>
Notes:
- Both runs and their saved files are revalidated. Results describe byte changes, not accessibility outcomes.
変更前後の検証済み保存証拠を、非公開レビュー用に比較します。
使用方法:
accessibility-audit compare-evidence --before <run.json> --after <run.json> --output <after-artifacts/new-private-comparison.json>
注意:
- Both runs and their saved files are revalidated. Results describe byte changes, not accessibility outcomes.
Create immutable private finding-management revisions linked to registered audit findings.
Usage:
accessibility-audit lifecycle init --run <run.json> --finding <finding-id> --output <artifacts/lifecycle-1.json> [--updated-at <RFC3339>]
accessibility-audit lifecycle advance --run <run.json> --before <prior-lifecycle.json> --input <state-patch.json> --output <artifacts/lifecycle-next.json>
Notes:
- Patch JSON may update status, assignee, accountable_owner, due_on, target_release, blocked_by, external_issues, decision, closure and updated_at.
- Each new file links the prior file hash. Accepted risk or exception never changes a human profile result into pass.
登録済みの監査findingに紐づく、不変で非公開のfinding管理revisionを作成します。
使用方法:
accessibility-audit lifecycle init --run <run.json> --finding <finding-id> --output <artifacts/lifecycle-1.json> [--updated-at <RFC3339>]
accessibility-audit lifecycle advance --run <run.json> --before <prior-lifecycle.json> --input <state-patch.json> --output <artifacts/lifecycle-next.json>
注意:
- Patch JSON may update status, assignee, accountable_owner, due_on, target_release, blocked_by, external_issues, decision, closure and updated_at.
- Each new file links the prior file hash. Accepted risk or exception never changes a human profile result into pass.
Read run state, evidence coverage, validation, successor warnings and next operations.
Usage:
accessibility-audit status --run <run.json> [--retest-of <predecessor-run.json>] [--lifecycle <latest-lifecycle.json> ...] [--as-of YYYY-MM-DD] [--format text|json] [--locale ja|en]
Options:
--run <run.json> Run manifest to inspect without modification.
--lifecycle <latest-record.json> Validate one finding-management chain; repeatable.
--as-of <YYYY-MM-DD> Date for overdue and exception checks; default is the host local date.
--format <text|json> Text or versioned JSON output. Default: text.
--locale <ja|en> Human-readable locale. JSON field names stay stable.
Notes:
- Read-only: no transition, target interaction or artifact write is performed.
- Successor discovery is limited to sibling JSON files. Report availability means a matching assessment can be rendered, not that inspection is complete.
監査の状態、証拠の範囲、検証結果、後続runと次の操作を読み取ります。
使用方法:
accessibility-audit status --run <run.json> [--retest-of <predecessor-run.json>] [--lifecycle <latest-lifecycle.json> ...] [--as-of YYYY-MM-DD] [--format text|json] [--locale ja|en]
オプション:
--run <run.json> 書き換えずに確認するrun。
--lifecycle <latest-record.json> Validate one finding-management chain; repeatable.
--as-of <YYYY-MM-DD> Date for overdue and exception checks; default is the host local date.
--format <text|json> textまたは版付きのjson。既定値: text。
--locale <ja|en> 表示言語。JSONのフィールド名は変わりません。
注意:
- 対象の操作、状態遷移、成果物の書込みは行いません。
- 後続runの検索は同じディレクトリのJSONに限ります。報告可能であることは検査完了を意味しません。
Create a new immutable audit run.
Usage:
accessibility-audit init --run-id <id> --profile <id> --inspection-mode <quick|detailed> --inspection-purpose <purpose> --target-name <name> --target-version <version> --target-ref <url|file> --artifact-root <directory> --network <none|local_read_only> --interaction <safe_read_only|human_supervised> --source-write <none|authorized_only> [--config <scope-environment.json>] --output <new-run.json>
Options:
--run-id <id> Unique run ID.
--profile <id> Active standards profile. Use `profiles list` to discover values.
--inspection-mode <quick|detailed> Choose a quick check or a detailed inspection for remediation before inspecting.
--inspection-purpose <purpose> Describe how the requester will use the report.
--target-name <name> Human-readable target name.
--target-version <version> Declared target version or commit.
--target-ref <url|file> Target URL or file. Repeat for additional target references.
--artifact-root <directory> Existing private directory for run artifacts.
--network <none|local_read_only> Network policy alias. `none` denies access; `local_read_only` maps to an allowlisted read policy.
--network-policy <policy.json> Required for allowlisted/local_read_only: explicit origins or exact URLs, GET/HEAD and redirect/resource policy. Propose with network-policy.
--interaction <safe_read_only|human_supervised> Interaction policy alias.
--interaction-policy <policy.json> Required for human_supervised: concrete operations, supervisor, scope, expiry and approval mode. A saved policy does not authorize runtime interaction.
--source-write <none|authorized_only> Source-write policy. Standard audits normally use `none`.
--config <file> Optional JSON file declaring scope and/or environment.
--supersedes-run <file> Validated predecessor required only for retest initialization.
--output <new-run.json> New output path; existing files are never overwritten.
Notes:
- The artifact root must already exist and must not traverse a symbolic link or reparse point.
- Inspection mode and purpose are required for new inspections. Retests inherit an existing request; changing it requires a separate inspection.
- Initialization records scope and permissions but does not inspect the target or create a profile outcome.
新しい不変のaudit runを作成します。
使用方法:
accessibility-audit init --run-id <id> --profile <id> --inspection-mode <quick|detailed> --inspection-purpose <purpose> --target-name <name> --target-version <version> --target-ref <url|file> --artifact-root <directory> --network <none|local_read_only> --interaction <safe_read_only|human_supervised> --source-write <none|authorized_only> [--config <scope-environment.json>] --output <new-run.json>
オプション:
--run-id <id> 一意のrun ID。
--profile <id> activeな規格profile。値は`profiles list`で確認できます。
--inspection-mode <quick|detailed> 検査前に、簡易チェック（quick）か詳細検査・改善用（detailed）を選びます。
--inspection-purpose <purpose> 依頼者がレポートを何に使うかを記録します。
--target-name <name> 人が読める対象名。
--target-version <version> 申告する対象版またはcommit。
--target-ref <url|file> 対象URLまたはfile。複数指定できます。
--artifact-root <directory> run artifactを保存する既存の非公開directory。
--network <none|local_read_only> network policy alias。`none`は拒否、`local_read_only`はallowlist付きread policyへ変換します。
--network-policy <policy.json> Required for allowlisted/local_read_only: explicit origins or exact URLs, GET/HEAD and redirect/resource policy. Propose with network-policy.
--interaction <safe_read_only|human_supervised> interaction policy alias。
--interaction-policy <policy.json> Required for human_supervised: concrete operations, supervisor, scope, expiry and approval mode. A saved policy does not authorize runtime interaction.
--source-write <none|authorized_only> source-write policy。通常の監査では`none`を使います。
--config <file> scopeまたはenvironmentを申告する任意のJSON file。
--supersedes-run <file> retest初期化時だけ必要な検証済みpredecessor。
--output <new-run.json> 新規出力path。既存fileは上書きしません。
注意:
- artifact rootは既に存在し、symbolic linkまたはreparse pointを経由してはいけません。
- 新規検査ではレベルと利用目的が必須です。再検査は既存の依頼内容を引き継ぎ、変更する場合は別の検査を開始します。
- 初期化はscopeとpermissionを記録するだけで、対象を検査したりprofile outcomeを作成したりしません。
Create a validator-valid assessment record or an explicit placeholder template.
Usage:
accessibility-audit assessment --profile <id> --target-name <name> --target-version <version> --target-ref <url|file> --evaluator <name> --evaluated-at <YYYY-MM-DD> --output <assessment.json>
accessibility-audit assessment --template --profile <id> --output <assessment.template.json>
Options:
--profile <id> Active standards profile.
--target-name <name> Required in record mode.
--target-version <version> Required in record mode.
--target-ref <url|file> Required in record mode; repeatable.
--evaluator <name> Required in record mode.
--evaluated-at <YYYY-MM-DD> Calendar date for the record.
--assessment-id <unique-id> Optional assessment identity. The default is a fresh UUID-based ID; never reuse an ID for another target or scope.
--template Create an editable placeholder, not a validated inspection record.
--output <file> New output path.
Notes:
- All profile rows begin as `not_tested`; creation is not a completed inspection.
validator-validなassessment recordまたは明示的なplaceholder templateを作成します。
使用方法:
accessibility-audit assessment --profile <id> --target-name <name> --target-version <version> --target-ref <url|file> --evaluator <name> --evaluated-at <YYYY-MM-DD> --output <assessment.json>
accessibility-audit assessment --template --profile <id> --output <assessment.template.json>
オプション:
--profile <id> activeな規格profile。
--target-name <name> record modeで必須。
--target-version <version> record modeで必須。
--target-ref <url|file> record modeで必須。複数指定できます。
--evaluator <name> record modeで必須。
--evaluated-at <YYYY-MM-DD> recordの暦日。
--assessment-id <unique-id> Optional assessment identity. The default is a fresh UUID-based ID; never reuse an ID for another target or scope.
--template 検証済み検査recordではなく編集用placeholderを作成します。
--output <file> 新規出力path。
注意:
- 全profile rowは`not_tested`で始まります。作成しただけでは検査完了ではありません。
Run rule-based browser checks and create a compact AI review context.
Usage:
accessibility-audit scan-web --url <http-or-https-url> --profile <active-profile> --output <new-scan.json> [--context-output <new-context.json>] [--allow-origin <origin>] [--allow-localhost] [--focus-steps <0-50>] [--width <240-7680>] [--height <240-7680>] [--reflow-width <240-1280>]
Options:
--url <URL> Public HTTP(S) target URL.
--profile <id> Active Web profile.
--output <file> Full internal scan artifact.
--context-output <file> Optional compact AI context.
--axe-output <file> Private native axe results with capture and per-frame execution metadata; requires --evidence-output.
--evidence-output <file> Private saved web bundle for capture-targets; requires --axe-output.
--browser-channel <chrome> Use installed system Chrome explicitly instead of the default browser runtime.
--allow-origin <origin> Additional explicit origin; repeatable.
--allow-url <exact-URL> Explicit exact URL including query; repeatable.
--run <run.json> Enforce a validated run policy as well as explicit caller grants. Requires --network-log-output.
--network-log-output <artifacts/network.json> Private per-request run/policy log. Saved on captured failures too. Cross-origin iframe capture is unverified and blocked.
--allow-localhost Permit loopback only for controlled fixtures.
--focus-steps <0-50> Keyboard focus steps; default 8.
--width <240-7680> Primary viewport width; default 1280.
--height <240-7680> Primary viewport height; default 800.
--reflow-width <240-1280> Reflow viewport width; default 320.
Defaults:
--focus-steps 8
--width 1280
--height 800
--reflow-width 320
Notes:
- Automated findings remain screening evidence and never become profile pass/fail outcomes automatically.
rule-based browser checkを実行し、圧縮したAI review contextを作成します。
使用方法:
accessibility-audit scan-web --url <http-or-https-url> --profile <active-profile> --output <new-scan.json> [--context-output <new-context.json>] [--allow-origin <origin>] [--allow-localhost] [--focus-steps <0-50>] [--width <240-7680>] [--height <240-7680>] [--reflow-width <240-1280>]
オプション:
--url <URL> 公開HTTP(S)対象URL。
--profile <id> activeなWeb profile。
--output <file> 完全な内部用scan artifact。
--context-output <file> 任意の圧縮AI context。
--axe-output <file> Private native axe results with capture and per-frame execution metadata; requires --evidence-output.
--evidence-output <file> Private saved web bundle for capture-targets; requires --axe-output.
--browser-channel <chrome> Use installed system Chrome explicitly instead of the default browser runtime.
--allow-origin <origin> 追加で明示するorigin。複数指定できます。
--allow-url <exact-URL> Explicit exact URL including query; repeatable.
--run <run.json> Enforce a validated run policy as well as explicit caller grants. Requires --network-log-output.
--network-log-output <artifacts/network.json> Private per-request run/policy log. Saved on captured failures too. Cross-origin iframe capture is unverified and blocked.
--allow-localhost 制御されたfixtureに限りloopbackを許可します。
--focus-steps <0-50> keyboard focus step数。既定値8。
--width <240-7680> 主viewport幅。既定値1280。
--height <240-7680> 主viewport高さ。既定値800。
--reflow-width <240-1280> reflow viewport幅。既定値320。
既定値:
--focus-steps 8
--width 1280
--height 800
--reflow-width 320
注意:
- 自動検出結果はscreening evidenceのままであり、profileのpass/failへ自動変換しません。
List active standards profiles, counts, groups, sources, and claim ceilings.
Usage:
accessibility-audit profiles list [--locale ja|en] [--format text|json|markdown]
Options:
--format <text|json|markdown> Output format; default text.
--locale <ja|en> Human-readable locale.
activeな規格profile、件数、group、source、claim上限を一覧表示します。
使用方法:
accessibility-audit profiles list [--locale ja|en] [--format text|json|markdown]
オプション:
--format <text|json|markdown> 出力形式。既定値はtext。
--locale <ja|en> 表示言語。既定値はen。
List, search, or show WCAG/JIS requirements without requiring internal IDs.
Usage:
accessibility-audit requirements list [--profile <id>] [--level A|AA] [--procedure available|unavailable] [--locale ja|en] [--format text|json|markdown]
accessibility-audit requirements search <query> [filters]
accessibility-audit requirements show <internal-id|success-criterion> [--profile <id>] [--locale ja|en] [--format text|json|markdown]
Options:
--profile <id> Limit results to one active profile.
--level <A|AA> Limit results by conformance level.
--procedure <available|unavailable> Filter criterion-specific human procedure availability.
--locale <ja|en> Preferred display language; default en.
--format <text|json|markdown> Output format; default text.
Notes:
- Search results are metadata and reproducibility aids, not conformance determinations.
内部IDを知らなくてもWCAG・JIS条項を一覧・検索・表示できます。
使用方法:
accessibility-audit requirements list [--profile <id>] [--level A|AA] [--procedure available|unavailable] [--locale ja|en] [--format text|json|markdown]
accessibility-audit requirements search <query> [filters]
accessibility-audit requirements show <internal-id|success-criterion> [--profile <id>] [--locale ja|en] [--format text|json|markdown]
オプション:
--profile <id> 1つのactive profileへ絞り込みます。
--level <A|AA> 適合レベルで絞り込みます。
--procedure <available|unavailable> 条項固有人手手順の有無で絞り込みます。
--locale <ja|en> 表示言語。既定値はen。
--format <text|json|markdown> 出力形式。既定値はtext。
注意:
- 検索結果はmetadataと再現性補助であり、適合性の判定ではありません。
Show one registered requirement by exact internal ID (legacy-compatible entry point).
Usage:
accessibility-audit requirement --profile <id> --id <requirement-id> [--locale ja|en] [--format json|markdown]
Options:
--profile <id> Active profile.
--id <requirement-id> Exact registered internal ID.
--format <json|markdown> Output format; default json.
--locale <ja|en> Human-readable locale.
Notes:
- For discovery by number or keyword, use `requirements show` or `requirements search`.
正確な内部IDで登録済み条項を1件表示します（legacy互換入口）。
使用方法:
accessibility-audit requirement --profile <id> --id <requirement-id> [--locale ja|en] [--format json|markdown]
オプション:
--profile <id> active profile。
--id <requirement-id> 登録済みの正確な内部ID。
--format <json|markdown> 出力形式。既定値はjson。
--locale <ja|en> 表示言語。既定値はen。
注意:
- 番号やkeywordから探す場合は`requirements show`または`requirements search`を使用します。
Measure Web inspection capabilities on an isolated fixture without inspecting a target.
Usage:
accessibility-audit preflight-web [--browser-channel chrome] [--require <comma-separated-capabilities>] [--format text|json]
Options:
--browser-channel <chrome> Use installed system Chrome; otherwise use the Playwright Chromium runtime.
--require <capabilities> Require a comma-separated subset of the six registered capabilities.
--format <text|json> Output format; default text.
--locale <ja|en> Human-readable locale.
Notes:
- Missing required capabilities exit with code 4. Screen-reader runtime remains externally unconfirmed. Success is not completed target inspection.
対象を検査せずに、分離したfixtureでWeb検査の実行能力を測定します。
使用方法:
accessibility-audit preflight-web [--browser-channel chrome] [--require <comma-separated-capabilities>] [--format text|json]
オプション:
--browser-channel <chrome> Use installed system Chrome; otherwise use the Playwright Chromium runtime.
--require <capabilities> Require a comma-separated subset of the six registered capabilities.
--format <text|json> Output format; default text.
--locale <ja|en> Human-readable locale.
注意:
- Missing required capabilities exit with code 4. Screen-reader runtime remains externally unconfirmed. Success is not completed target inspection.
Diagnose Node, package, registry, distribution, and optional browser capabilities.
Usage:
accessibility-audit doctor [--locale ja|en] [--format text|json]
Options:
--format <text|json> Output format; default text.
--locale <ja|en> Human-readable locale.
Notes:
- The command is read-only. Missing optional browser capability produces WARN rather than a false package failure. Run preflight-web to test the actual browser runtime.
Node、package、registry、distribution、任意browser capabilityを診断します。
使用方法:
accessibility-audit doctor [--locale ja|en] [--format text|json]
オプション:
--format <text|json> 出力形式。既定値はtext。
--locale <ja|en> 表示言語。既定値はen。
注意:
- read-only commandです。任意browser capabilityの欠落はpackage failureではなくWARNとして示します。
Create, validate, report, and compare private participation reviews for non-Web targets.
Usage:
accessibility-audit non-web-review init --kind <document-slide|media-content|event-community|participation-workflow> --id <id> --name <name> --version <version> --scope <scope> --output <new.json>
accessibility-audit non-web-review validate --input <review.json>
accessibility-audit non-web-review report --input <review.json> --output <new.md>
accessibility-audit non-web-review compare --before <old.json> --after <new.json> --output <new.md|json> [--format markdown|json] [--locale ja|en]
Options:
--locale <ja|en> Template questions and report language; default ja.
--format <markdown|json> Comparison output format; default markdown.
Notes:
- No Web audit run or standards profile is created. Reports and comparisons are private participation records, not conformance claims.
Web以外の対象について、非公開の参加レビューを作成・検証・報告・比較します。
使用方法:
accessibility-audit non-web-review init --kind <document-slide|media-content|event-community|participation-workflow> --id <id> --name <name> --version <version> --scope <scope> --output <new.json>
accessibility-audit non-web-review validate --input <review.json>
accessibility-audit non-web-review report --input <review.json> --output <new.md>
accessibility-audit non-web-review compare --before <old.json> --after <new.json> --output <new.md|json> [--format markdown|json] [--locale ja|en]
オプション:
--locale <ja|en> Template questions and report language; default ja.
--format <markdown|json> Comparison output format; default markdown.
注意:
- No Web audit run or standards profile is created. Reports and comparisons are private participation records, not conformance claims.
Show supporting checks for stateful UI and screen-reader behavior.
Usage:
accessibility-audit screen-reader-checklist [--list-patterns] [--pattern id|all] [--extension file.json] [--locale ja|en] [--format json|markdown]
Options:
--list-patterns List available bundled and extension patterns.
--pattern <id|all> Checklist pattern; default all.
--extension <file.json> Optional JSON file with additional patterns.
--format <json|markdown> Output format; default json.
--locale <ja|en> Human-readable locale.
状態を持つUIとscreen-reader動作の補助checkを表示します。
使用方法:
accessibility-audit screen-reader-checklist [--list-patterns] [--pattern id|all] [--extension file.json] [--locale ja|en] [--format json|markdown]
オプション:
--list-patterns List available bundled and extension patterns.
--pattern <id|all> checklist pattern。既定値はall。
--extension <file.json> Optional JSON file with additional patterns.
--format <json|markdown> 出力形式。既定値はjson。
--locale <ja|en> 表示言語。既定値はen。
Validate an immutable audit run and write a new validation record.
Usage:
accessibility-audit validate-run --input <run.json> --output <new-validation.json> [--historical-resources <saved-skill-root>]
Options:
--input <run.json> Audit run.
--output <file> New validation record.
--historical-resources <skill-root> Saved resource files whose hashes match this run; read-only.
不変のaudit runを検証し、新しいvalidation recordを書き出します。
使用方法:
accessibility-audit validate-run --input <run.json> --output <new-validation.json> [--historical-resources <saved-skill-root>]
オプション:
--input <run.json> audit run。
--output <file> 新しいvalidation record。
--historical-resources <skill-root> Saved resource files whose hashes match this run; read-only.
Validate an assessment and print its coverage and claim guard result.
Usage:
accessibility-audit validate-assessment <assessment.json> [--run <run.json>] [--trust-policy <recipient-selected.json> --trust-policy-sha256 <independent canonical SHA-256>]
Notes:
- Current run-backed review records require --run and its original artifacts. Stored authenticated flags are never accepted.
assessmentを検証し、coverageとclaim guard結果を表示します。
使用方法:
accessibility-audit validate-assessment <assessment.json> [--run <run.json>] [--trust-policy <recipient-selected.json> --trust-policy-sha256 <independent canonical SHA-256>]
注意:
- Current run-backed review records require --run and its original artifacts. Stored authenticated flags are never accepted.
Register one validated artifact in a new audit-run version.
Usage:
accessibility-audit register --run <run.json> --artifact <artifact.json> [--allow-origin <origin>] [--allow-url <exact-URL>] [--allow-localhost true] [--network-log-output <artifacts/network.json>] --output <new-run.json>
Options:
--run <run.json> Current run version.
--artifact <artifact.json> Validated artifact within its artifact root.
--allow-origin <origin> Explicit HTTP remeasurement origin; repeatable.
--allow-url <exact-URL> Explicit HTTP remeasurement URL; repeatable.
--network-log-output <artifacts/network.json> Required for HTTP remeasurement; private run-bound request log.
--allow-localhost <true> Permit loopback only for controlled HTTP fixtures.
--output <new-run.json> New immutable run version.
検証済みartifactを新しいaudit-run版へ1件登録します。
使用方法:
accessibility-audit register --run <run.json> --artifact <artifact.json> [--allow-origin <origin>] [--allow-url <exact-URL>] [--allow-localhost true] [--network-log-output <artifacts/network.json>] --output <new-run.json>
オプション:
--run <run.json> 現在のrun版。
--artifact <artifact.json> artifact root内の検証済みartifact。
--allow-origin <origin> Explicit HTTP remeasurement origin; repeatable.
--allow-url <exact-URL> Explicit HTTP remeasurement URL; repeatable.
--network-log-output <artifacts/network.json> Required for HTTP remeasurement; private run-bound request log.
--allow-localhost <true> Permit loopback only for controlled HTTP fixtures.
--output <new-run.json> 新しい不変run版。
Merge registered artifacts into a new assessment.
Usage:
accessibility-audit merge --run <run.json> --assessment <assessment.json> --artifact <artifact.json> [--claim-tier <reference_only|screened|evaluated_subset>] --output <new-assessment.json>
Options:
--run <run.json> Validated current run.
--assessment <assessment.json> Baseline assessment.
--artifact <artifact.json> Registered artifact; repeatable.
--review-record <record.json> Portable review record bound to a registered human artifact; repeatable.
--trust-policy <recipient-selected.json> External recipient policy; requires --trust-policy-sha256.
--trust-policy-sha256 <canonical SHA-256> Policy pin obtained independently of the audit bundle.
--claim-tier <reference_only|screened|evaluated_subset> Explicit claim request checked against registered evidence. Default: reference_only. Wording comes from the registry.
--output <new-assessment.json> New merged assessment.
登録済みartifactを新しいassessmentへ統合します。
使用方法:
accessibility-audit merge --run <run.json> --assessment <assessment.json> --artifact <artifact.json> [--claim-tier <reference_only|screened|evaluated_subset>] --output <new-assessment.json>
オプション:
--run <run.json> 検証済みの現在run。
--assessment <assessment.json> baseline assessment。
--artifact <artifact.json> 登録済みartifact。複数指定できます。
--review-record <record.json> Portable review record bound to a registered human artifact; repeatable.
--trust-policy <recipient-selected.json> External recipient policy; requires --trust-policy-sha256.
--trust-policy-sha256 <canonical SHA-256> Policy pin obtained independently of the audit bundle.
--claim-tier <reference_only|screened|evaluated_subset> 登録済み証拠で検証する主張範囲。既定値: reference_only。文言はregistryの固定テンプレートです。
--output <new-assessment.json> 新しいmerged assessment。
Render a profile-aware, provenance-explicit Markdown report from a validated standalone or run-backed assessment.
Usage:
accessibility-audit report --input <assessment.json> [report options]
accessibility-audit report --run <audit-run.json> --assessment <assessment.json> --output <new-report.md|html> [--historical-resources <saved-skill-root>] [report options]
Options:
--input <assessment.json> Standalone interface.
--run <audit-run.json> Run-backed interface; requires --assessment and --output.
--assessment <assessment.json> Merged run-backed assessment.
--historical-resources <skill-root> Use saved resources matching the run hashes for a read-only historical report.
--lifecycle <latest-lifecycle.json> Add a validated private lifecycle status summary for a registered finding; repeatable.
--as-of <YYYY-MM-DD> Date for overdue and exception-expiry checks; default is the host local date.
--trust-policy <recipient-selected.json> Reverify reviewer signatures under an external policy; requires its independent pin.
--trust-policy-sha256 <canonical SHA-256> Independent policy pin. Omitting both trust flags leaves valid signatures self-signed.
--locale <ja|en> Human-readable report locale; default ja. IDs and enum values do not change.
--output <report.md> New report path. Standalone mode may write to stdout when omitted.
--format <markdown|html> Output format. Default: markdown.
--detail <summary|full> Summary or complete 55/56-row report. Default: full.
--appendix <file> With summary, write the complete report in the selected format.
--visibility <internal|public> Publication boundary. Default: internal.
--reviewer-disclosure <include|redact> Required for public output.
--redaction-manifest <manifest.json> Required internal record for public output.
Notes:
- Use either --input or --run/--assessment, never both.
- Screening projections are report-only judgements and are never promoted to human-verified profile outcomes.
- Existing files are never overwritten.
検証済みstandaloneまたはrun-backed assessmentから、profile対応・provenance明示のレポートを生成します。
使用方法:
accessibility-audit report --input <assessment.json> [report options]
accessibility-audit report --run <audit-run.json> --assessment <assessment.json> --output <new-report.md|html> [--historical-resources <saved-skill-root>] [report options]
オプション:
--input <assessment.json> standalone interface。
--run <audit-run.json> run-backed interface。--assessmentと--outputが必要です。
--assessment <assessment.json> 統合済みrun-backed assessment。
--historical-resources <skill-root> Use saved resources matching the run hashes for a read-only historical report.
--lifecycle <latest-lifecycle.json> Add a validated private lifecycle status summary for a registered finding; repeatable.
--as-of <YYYY-MM-DD> Date for overdue and exception-expiry checks; default is the host local date.
--trust-policy <recipient-selected.json> Reverify reviewer signatures under an external policy; requires its independent pin.
--trust-policy-sha256 <canonical SHA-256> Independent policy pin. Omitting both trust flags leaves valid signatures self-signed.
--locale <ja|en> 人向けレポートの言語。既定値はja。IDとenum値は変わりません。
--output <report.md> 新規レポートpath。standaloneでは省略時にstdoutへ出力できます。
--format <markdown|html> 出力形式。markdownまたはhtml。既定値はmarkdown。
--detail <summary|full> summaryまたは全55／56行を含むfull。既定値はfull。
--appendix <file> summaryと同じ形式で完全版付録を書き出します。
--visibility <internal|public> internalまたはpublic。既定値はinternal。
--reviewer-disclosure <include|redact> public出力時のreviewer表示方針。
--redaction-manifest <manifest.json> public出力時に必要な内部用伏字記録。
注意:
- --inputと--run/--assessmentは同時に使用できません。
- screening projectionはreport-only judgementであり、人手確認済みprofile outcomeへ昇格しません。
- 既存fileは上書きしません。
- 自動伏字は公開承認ではありません。公開前の人による確認が必要です。
Create a fresh audit run from a registered authorized or externally declared change.
Usage:
accessibility-audit retest --supersedes-run <old-run.json> [all init options for the new target version]
Notes:
- Prior evidence and outcomes are not silently inherited.
認可済み変更が完了したpredecessorから新しいaudit runを作成します。
使用方法:
accessibility-audit retest --supersedes-run <old-run.json> [all init options for the new target version]
注意:
- 以前の証拠とoutcomeを黙って継承しません。