Candidate implementation, not yet complete with actual host, native installation and creative acceptance. Only macOS arm64, Python 3.11+.
Routing scope
Use only when the user requests first-time installation, checking installation status, or troubleshooting pinned CLI installation failures. Ordinary typesetting work goes to designcraft-use(Install: npx skills add full-aigc-skills/designcraft-skills --skill designcraft-use); querying native commands goes to designcraft-cli(Install: npx skills add full-aigc-skills/designcraft-skills --skill designcraft-cli). Installation will access the network to obtain pinned artifacts and write to the user-selected runtime-home; authorization covering that operation must be obtained first; checking only does not trigger installation. On failure, retain logs and existing files, and do not automatically retry writes. Acceptance is based on pinned digests, version checks, and actual command discovery records; successful installation does not mean the host has discovered the skill or that the creative workflow has passed.
When to Use
Use when the user explicitly requests checking whether the DesignCraft pinned CLI is installed, installing the runtime for the first time, or troubleshooting installation failures. Ordinary queries go to designcraft-cli, and ordinary editing requests go to designcraft-use。
Out of scope and security boundaries
- A read-only check is not installation permission; network downloads and installation require explicit authorization for this dependency.
- Do not use sudo, do not modify PATH, do not log in to accounts, and do not upgrade to artifacts outside the locked version.
- The target runtime-home must be selected by the user or follow an approved default directory, and old files that cannot be verified must never be overwritten.
- Runtime installation, host discovery, model routing, and native creation are accepted separately.
Workflow
Step 1: Confirm the platform and target directory
Record the macOS architecture, Python version, and isolated runtime-home; do not attempt installation on unsupported platforms.
Step 2: Perform a read-only check first
Inspect the lock file, target directory, and installed identity; existing artifacts must simultaneously match the digest, version, and installation receipt.
Step 3: Verify installation authorization
Before installation, explicitly verify whether the user has authorized this network download and file write. Without authorization, only report the dependency gap.
Step 4: Verify and install the locked artifacts
Fetch only the artifact corresponding to the locked URL, and verify the archive and binary digests; on installation failure, retain the error classification, do not overwrite abnormal directories, and do not automatically retry.
Step 5: Layered verification
Verify the installation receipt, version output, and command discovery records; separately mark the status of plugin discovery, model routing, and creative acceptance.
Rules (execution constraints)
- This skill will not interpret the user's typesetting goal as authorization to download the runtime.
- If the digest or installation receipt does not match, retain the directory for diagnosis and do not silently reinstall.
- Successful installation only proves a local identity check of the locked binary, not host or business capabilities.
Gotchas (common misconceptions)
- When the platform does not match, do not attempt a compatible installation: report the current supported scope and the missing evidence.
- An available download link does not mean the artifact is trustworthy: verify the archive digest and binary digest every time.
- An existing directory does not mean installation passed: verify the installation receipt and version output.
- After an installation failure, do not automatically retry and overwrite: preserve logs and directory state for manual verification.
- Passing installation does not mean Codex has loaded it: The host discovery requires separate verification.
First use
Set SKILL_DIR to the absolute directory where the SKILL.md actually loaded by the host resides.
The skill can be installed independently into the project .agents/skills, the user skills directory, or the plugin cache; do not guess sibling directories, and do not modify skill files.
When the user request has authorized the necessary dependencies, the public cli.py automatically verifies and installs the pinned native version into the user data directory; it does not use sudo and does not modify PATH.
: "${SKILL_DIR:?设置为本技能真实加载目录}"
python3 -I -B "$SKILL_DIR/scripts/cli.py" -- commands
When you only need to check the installation:
python3 -I -B "$SKILL_DIR/scripts/bootstrap.py"
python3 -I -B "$SKILL_DIR/scripts/cli.py" -- --version
--runtime-home or CRAFT_RUNTIME_HOME specifies an isolated runtime directory, placed before -- .
--archive points to an existing pinned ZIP, and the archive and binary digests are still verified.
On installation failure, dependencySetup is returned; do not automatically retry editing or rendering; on verification failure, keep the existing directory.
The current lock only pins the native CLI; the release snapshot and host auto-discovery still require separate verification.
Current scenario
First record the user's assets, goals, output location, dimensions/page count/format, and modification boundaries. Query the real parameters from the native directory, confirm the command, then execute; do not invent commands, and do not treat new interfaces in local source code as already supported by the pinned artifact. The asset paths and output locations in the examples below must be replaced with the paths actually authorized for the current task.
python3 -I -B "$SKILL_DIR/scripts/cli.py" -- --version
Deliver the native .designcraft project, the collected asset and font license inventory, page-by-page previews, and applicable PDF/IDML/EPUB. Reopen the native project to verify page count, text, and linked assets; record interchange format losses separately.
CLI arguments are passed via argv and do not execute a shell. Native commands may have file-write and network behavior; use them only within the scope authorized by the task. On execution failure, first check the project and output; do not replay writes for unknown results. For output to the same path, protect existing files first. A command query or zero exit does not mean delivery passed: check the actual files, save and reopen, and the content and modified results, and also record a creative review. The runtime and skill installation, host skill discovery, and model natural-language selection are gates verified separately.
Scenario examples
The post-installation verification success path of this skill and the rejection and recovery boundaries when authorization is missing or installation fails are in Scenario examples、read-only check、authorized installationandverification failure. The examples only illustrate the process and do not prove that the pinned native CLI has been accepted.
Required reading before execution
Before executing a command plan, checkpoint recovery, or delivery verification, you must read this skill's execution and acceptance details, and operate according to its input preservation, UNKNOWN, receipt, and format verification requirements. Digests can only detect changes and cannot isolate or roll back writes; unknown results must not be automatically replayed.