This skill handles querying actual command parameters and capabilities, and invoking the public CLI, MCP, and diagnostics. Hand off to skills in the same package by name; it can be used when installed alone and does not read sibling directories. Invoke the fixed released filmcraft-cli and preserve the native editing project.
Trusted directory permissions
Legacy cli.py Native-forwarded project reading and editing also require an independent root parameter before -- ; only exact version/help/no-project-directory queries remain compatible with legacy invocation. Model downloads must separately use --model-maintenance, see this skill's Execution permissions。
Public workflow.py and commands.py run must explicitly provide --read-root READ_DIRECTORY --write-root WRITE_DIRECTORY(repeatable); the root must be an existing canonical real directory, authorized by the user/host, and must not be derived from plans or asset metadata. Plans, assets, and source deliverables must be under the read root, and deliverables and their parent directories must be under the write root.--runtime-home is the installation maintenance directory and must be under a separate, explicitly authorized write root; the native editing process is always prohibited from modifying that cache, skills, registration inputs, and source projects. Current root-isolated execution supports macOS headless; the desktop entry must simultaneously isolate the signed application owned by this session and the MCP client, exposing only the allocated loopback control port. External bridges whose isolation cannot be confirmed are explicitly rejected. The low-level Python testing API is not a public authorized entry point. For detailed boundaries see Execution permissions。
Inputs and deliverables
Inputs are the user-confirmed task, assets, project or object, output directory, and modification scope; when an existing project is required, first verify the summary. Return the actual CLI results, the saved native project, required derived outputs, and verification records. Report installation success, command directory existence, and creative task completion separately.
First use and public entry points
Locate the real directory of the current SKILL.md. Currently supports macOS arm64, Python 3.11+; the fixed CLI is installed to the user data directory. When an existing task authorization covers the necessary dependencies, directly execute this skill's installer without creating a separate approval process.
Set SKILL_DIR to the directory (absolute path) of this SKILL.md actually loaded by the host. User-level installations may be located at ~/.agents/skills/filmcraft-cli, project-level ones may be located at .agents/skills/filmcraft-cli, and plugin ones may be located in their skills/filmcraft-cli or 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 arbitrarily choose among duplicate versions. The skill directory and the CLI user data installation directory are two independent locations.
: "${SKILL_DIR:?请先设置为本 SKILL.md 的实际所在目录}"
python3 -I -B "$SKILL_DIR/scripts/bootstrap.py"
python3 -I -B "$SKILL_DIR/scripts/cli.py" -- --version
python3 -I -B "$SKILL_DIR/scripts/cli.py" -- commands --json
CLI argv is after -- , and the native subcommand must come first. Installation parameters go before the separator;--runtime-home can isolate the cache. Stop when locked artifact digest verification 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 describes startup parameters.
Scenario operations
First query the current version and command directory; record read-only and editing invocations separately. After reading the current project state and command enabled/params, construct argv; consecutive steps that need to create objects must be executed in the same session, and returned IDs must not be guessed.
describe <id> check the parameters, then use exec <id> <JSON> or run <JSONL>; write commands must be saved in the same batch or used together with --save-as。
Native composition and source project revision use this skill's bundled scripts/workflow.py; read Workflow contract, the template is in this skill's examples. Only modify the authorized object; native projects and dependency assets are retained, and derived formats lose read Exchange report。
Verification and recovery
A non-zero command exit does not count as completion; you need to inspect the actually saved project, exports, dimensions, and media decoding. Timed-out results are marked unknown; first check the original task/project, and do not automatically replay edits. Existing native workflows support save-as revision; when the corresponding project summary does not match, refuse to overwrite.
Reference and handoff as needed
- Actual command evidence: fixed-version observation, used only as routing and parameter reference; real-time results take precedence, and disabled items are not executed.
- Workflow contract: combined operations, native saving, dependency collection, and revision.
- Hand off to filmcraft-cli-setupwhen installation/diagnostics require it, and hand off complete task routing to filmcraft-use; when a skill is missing, use
npx skills add full-aigc-skills/filmcraft-skills --skill <skill-name>. Do not load other skills through adjacent file paths.
This skill does not provide a fictitious login interface; local headless does not require a cloud account. Full GUI, cross-editor fidelity, and creative quality are stated based on actual evidence.
Complete native command usage
The current skill includes the parameter descriptions and same-session entry for the complete catalog, and is not limited by the creative template whitelist. Read Complete usage guide, and query Command referenceas needed; each instruction has a skill route, prerequisite observation, and acceptance status.
: "${READ_ROOT:?Set READ_ROOT to an existing independently authorized directory}"
: "${RUNTIME_HOME:?Set RUNTIME_HOME to an existing independently authorized directory}"
: "${SKILL_DIR:?Set SKILL_DIR to an existing independently authorized directory}"
: "${WRITE_ROOT:?Set WRITE_ROOT to an existing independently authorized directory}"
python3 -I -B "$SKILL_DIR/scripts/commands.py" list --filter QUERY
python3 -I -B "$SKILL_DIR/scripts/commands.py" describe COMMAND_ID
python3 -I -B "$SKILL_DIR/scripts/commands.py" run "$SKILL_DIR/examples/commands-advanced.json" --output /absolute/new-command-result --runtime-home "$RUNTIME_HOME" --read-root "$SKILL_DIR" --read-root "$READ_ROOT" --write-root "$WRITE_ROOT" --write-root "$RUNTIME_HOME"
Before executing a new entry point, check the actual registry and current executable status, and retain return value references and step-by-step receipts; semantic errors or timeouts are not passed off as success. Catalog coverage and direct native use do not mean that all instructions, GUI, delivery, or Art orchestration have been accepted.
For the complete workflow command gateway, see Usage instructions. The version and artifact digest of the fixed CLI are based on this skill's included scripts/runtime.lock.json ; the skill package version is based on the corresponding release tag. Independent installation re-verification and full command-by-command acceptance are recorded separately, and release is not used as a substitute for acceptance.
For GUI tasks, you can first use this skill's included Fixed desktop installation; installation, launch, and actual GUI editing are verified separately.
Complete attribution commands are listed by command family in Categorized usage list, and each item can be viewed through this skill's commands.py describe for parameters; the general CLI list also retains all commands not assigned to specialized skills.
For this skill's tasks, first read Specialized task template, and execute according to the corresponding input, first task, failure boundaries, local revision, and verification. When combining across capabilities, then read Complete business scenarios。
If installation fails, read First-use diagnostics, and use the receipt to locate the current skill's own setup entry; handle installation failure and native invocation failure separately.
For competition for the same target execution and interruption registration, see this skill's Execution registration; the skill source dev.29 includes the implementation, and plugin fixed installation acceptance is recorded separately.
For continuous operations, prefer this skill's included Task-level session: start once, check, save, and rework plan by plan, sharing the same FilmCraft instance; exit only when the task ends. A single run still retains the original compatible behavior.