ClawHub

Task Director

v1.0.0

Task Director — turn complex tasks into movie storyboards. Create a plan, review it, then execute step by step with fallback support. Pause, retry, skip anyt...

⭐ 0· 43·0 current·0 all-time
by@besty0121
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Benign
medium confidence
✓
Purpose & Capability
The name/description (turn tasks into storyboards and orchestrate execution) matches the included files: a Python CLI (scripts/director.py), SKILL.md describing CLI usage, and example templates. Requiring only python is proportionate.
ℹ
Instruction Scope
SKILL.md instructs the agent to: call the CLI to get the next 'shot', then execute the shot.command in the environment and report back with director.py result commands. This is coherent for an orchestrator, but it means the agent will run arbitrary shell commands defined in scenes — the skill does not itself restrict or sanitize those commands. Users should review scene files before approving/executing them.
✓
Install Mechanism
No install spec — instruction-only with an included Python script. Nothing is downloaded from external URLs or written to nonstandard system locations beyond the user's home directory (~/.openclaw).
✓
Credentials
The skill declares no required environment variables or credentials. Data is stored locally under ~/.openclaw/memory/movies. No unrelated credentials, tokens, or config paths are requested.
✓
Persistence & Privilege
always is false and the skill's persistence is limited to per-user files in ~/.openclaw/memory/movies. It does not request elevated privileges or modify other skills' configs based on the provided files.
Assessment
This skill is internally coherent and appears to do what it says, but it delegates command execution to the agent: scene 'shots' can contain arbitrary shell commands (install packages, delete files, curl URLs, etc.). Before approving or running a movie, inspect the scenes.json (or any auto-generated scenes) and the director.py file for unexpected behavior. Prefer running this skill in a low-privilege or sandboxed environment if you will execute untrusted plans. If you rely on autonomous invocation, be aware an agent that has permission to run shell commands plus this orchestrator can perform system actions defined in scenes — review and limit what scenes are allowed.

Like a lobster shell, security has layers — review code before you run it.

latestvk976vf350qnvjd6errm2v1eyzn842b8s

License

MIT-0
Free to use, modify, and redistribute. No attribution required.
Termshttps://spdx.org/licenses/MIT-0.html

Runtime requirements

Binspython

Comments