This skill is responsible for checking the technical evidence of mixed deliverables and the consistency of visual copy and audio. It hands off to skills in the same package by name, can be used when installed alone, and does not read sibling directories. Use this project's local orchestration CLI; the upstream ArtCraft is a Tauri application, and no official CLI of the same name is installed or impersonated.
Execution implementation is consolidated in the built-in artcraft-harness skill (installed bundled with the plugin; for standalone installation use npx skills add full-aigc-skills/artcraft-skills --skill artcraft-harness). Set HARNESS_DIR to the directory where that skill is actually loaded (absolute path); this skill's SKILL_DIR is only used to read its own examples/ and references/, and script execution paths always use $HARNESS_DIR/scripts/。
Inputs and Deliverables
The inputs are the user-confirmed task, materials, project or object, output directory, and modification scope; when an existing project is needed, verify the summary first. Return the actual CLI results, the saved native project, the required derived outputs, and the verification records. Report installation success, command directory existence, and creative task completion separately.
First-Time Use and Public Entry Point
Locate the real directory of the current SKILL.md. Currently supports macOS arm64 and Python 3.11+; the base entry point installs the pinned Node and orchestration package; workflow.py installs the required domain skill sources and native CLI according to the task graph, and does not require global Node. When existing task authorization covers the necessary dependencies, execute this skill's installer directly and do not create a separate approval process.
Set SKILL_DIR to the directory (absolute path) where this SKILL.md actually loaded by the host resides. User-level installations may be located at ~/.agents/skills/artcraft-cli-review, project-level ones may be located at .agents/skills/artcraft-cli-review, and plugins may be located in their skills/artcraft-cli-review or the host cache directory; use the actual loaded path as the standard, do not guess based on the current working directory, and do not search and then arbitrarily select a duplicate version. The skill directory and the CLI's user data installation directory are two independent locations.
: "${SKILL_DIR:?请先设置为本 SKILL.md 的实际所在目录}"
python3 -I -B "$HARNESS_DIR/scripts/bootstrap.py" --runtime-only
python3 -I -B "$HARNESS_DIR/scripts/cli.py" -- --version
python3 -I -B "$HARNESS_DIR/scripts/cli.py" -- --help
In CLI argv, after -- , the native subcommand must come first. Installation parameters go before the separator;--runtime-home can isolate the cache. Stop when locking the artifact digest fails, the installation is corrupted, or the platform is unsupported; do not modify PATH and do not perform floating upgrades. The native help entry point is --help; the launcher's own --help only explains the startup parameters.
Scenario Operations
Verify this scenario's inputs, native project, target object, version, and output boundaries. According to the scenario guide , select the current command, save an independent checkpoint, and then execute; after completion, reopen the native project and check the actual output and non-target content.
This project's --help defines run/status/cancel/package/verify-package; semantic plans go through workflow.py, and the upstream Tauri's account/vendor methods must not be treated as orchestration CLI commands.
Technical SHA verification and actual visual and auditory review are kept separate; record specific issues, rework scope, and unverified items, and do not fabricate model review results.
Native composition and source project revisions use the artcraft-harness skill's scripts/workflow.py; read the workflow contract, and the templates are within this skill's examples. Modify only authorized objects, retain the native project and dependent materials, and for derived format loss read the exchange report。
Verification and Recovery
A non-zero command exit does not count as completion; the actual saved project, export, dimensions, and media decoding must be inspected. Timeout results are marked unknown; first check the original task/project, and do not automatically replay edits. Perform recovery and packaging according to this skill's references/recovery.md and references/project-package.md respectively; review_ready only indicates technical readiness.
On-demand references and handoff
- Actual command evidence: fixed-version observations, used only as routing and parameter references; real-time results take precedence, and disabled items are not executed.
- Workflow contract: composite operations, native saving, dependency collection, and revision.
- Hand off to artcraft-cli-setupwhen installation/diagnostics require it, and hand off full task routing to artcraft-use; when the skill is missing, use
npx skills add full-aigc-skills/artcraft-skills --skill <skill-name>. Do not load other skills via adjacent file paths.
This skill does not provide a fabricated login interface; local headless does not require a cloud account. Full GUI, cross-editor fidelity, and creative quality are stated according to actual evidence.
When the finished video needs to be verified via the existing Video Factory, read this skill's public verification handoff. This adaptation preserves NOT_RUN and does not replace creative review or the native project.
Save named review results
When technical, creative, human acceptance, and brand issues need to be saved, call the artcraft-harness skill's current delivery review record according to scripts/review.py record/verify. Use the current packaging receipt summary and actual observation files; bind child nodes, asset versions, and issue locations. Record only observations that were actually performed, do not fabricate human acceptance, and do not treat test records or model opinions as user consent. When a complete review is missing, keep it pending, and the ledger status remains review_ready.
python3 -I -B "$HARNESS_DIR/scripts/review.py" record --package "$PACKAGE_ROOT" --package-sha "$PACKAGE_SHA" --input "$REVIEW_INPUT" --output "$REVIEW_ROOT"
python3 -I -B "$HARNESS_DIR/scripts/review.py" verify --package "$PACKAGE_ROOT" --package-sha "$PACKAGE_SHA" --review "$REVIEW_ROOT" --review-sha "$REVIEW_SHA"
Dynamic transparent sequence handoff
When a hybrid task requires transparent animation, a complete frame handoff, or downstream updates after a logo change, use this skill'sdynamic sequence guideand examples/dynamic-brand-campaign.json. Verify all frames, time base, native references, and local rework according to actual receipts; do not substitute the first frame or ordinary JSON for the sequence.
When the result is unknown after saving, verify according to this skill's references/recovery.md , failure.json, the original temporary project, and stop evidence; do not replay the same task, and a failed temporary save is not a successful delivery.
Complete domain command handoff
When native operations beyond the templates are needed, read this skill'scomplete domain command component, use the included domain_commands.py to query 2646 commands, actual MCP parameters, and hand off the selected domain. Each skill includes creation/rework examples for the four domains; do not replay unknown, and component receipts do not replace native DAG delivery and acceptance.
Native gradients, multiple fills, and source project recoloring rework: use this skill's Vector appearance guide。
Native parenting, transparency expressions, and source project rework: use this skill's Effect expression guide。
For Photo native adjustment layers, selection masks, and trusted source rework, see local adjustment guide。
For independently launching domain GUI commands, see own desktop handoff。
Hybrid task reference business scenario manual, first confirm dependencies and input versions, then execute sub-projects, local rework, and delivery verification.
On installation failure, read first-use diagnostics, locate the setup entry point of the current skill itself according to the receipt; handle installation failures and native invocation failures separately.
For real voiceover recognition and mixed subtitle tasks, please read this skill's Mixed ASR guide, and execute according to the fixed domain dispatch and persistent model directory; for planning, rework, and acceptance, retain dependencies, native engineering, and execution evidence respectively.
Read the task identity and rejection details according to the task receipt description ; node summaries do not replace actual execution receipts.
Before reworking mixed posters, covers, and source projects, read Photo complete delivery, and distinguish between host-independent plugins and Art-internal fixed dependencies.
For cross-artifact observation of fixed brand and subject references, see this skill's consistency review contract; report acceptance of source candidates and published fixed versions separately.
Runtime upgrades and legacy ledgers
When encountering runtime_upgrade_busy , retain the old runtime, original project, and ledger; do not delete leases or forcibly modify state. First complete or reliably stop the task with the old version, then migrate according to the new version; the old budget marker of read-only status indicates historically untracked. See this skill'supgrade guide。
For validation and mode boundaries of native command execution, see command contract boundaries。
For classification, parameter differences, and rejection boundaries of mode directory candidates, seemode directory usage; the Vector desktop/bridge duplicate command has not yet completed execution qualification and cannot fall back according to the headless directory.