output-validation
v0.1.0Local self-check of instructions and mask outputs (format/range/consistency) without using GT.
⭐ 0· 32·0 current·0 all-time
by@wu-uk
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name and description describe local output validation; SKILL.md contains focused checks (key format, coverage, NPZ/CSR structure, value validity) that directly implement that purpose. Nothing in the instructions attempts to access unrelated services or secrets.
Instruction Scope
Instructions operate on user-supplied local files (VIDEO_PATH, INSTRUCTIONS_PATH, MASKS_PATH) and perform file- and content-level assertions. They do not ask the agent to read unrelated system files, environment variables, or send data externally. The scope stays within validating generated outputs.
Install Mechanism
This is instruction-only (no install spec), which is low-risk. The reference snippet uses Python libraries (numpy, cv2) but the skill does not declare runtime dependencies or required binaries; users should ensure Python + numpy + OpenCV are available in their environment.
Credentials
No environment variables, credentials, or config paths are requested. The skill's data access is limited to the file paths the user supplies, which is proportionate to its validation purpose.
Persistence & Privilege
The skill is not always-on and does not request persistent or elevated privileges. It does not modify other skills or system-wide settings.
Assessment
This skill is an instruction-only local validator for interval/mask outputs and appears coherent with that purpose. Before installing/using: (1) ensure you run it in an environment with Python, numpy, and OpenCV (cv2) available, since the snippet assumes those libraries; (2) provide only trusted local VIDEO_PATH / INSTRUCTIONS_PATH / MASKS_PATH files (the checks read those files but do not contact external services); (3) be prepared for the snippet to raise assertions on invalid data — consider running it in an isolated environment or CI step; and (4) if you need automated installation, request or add an explicit install spec declaring required packages to avoid surprises.Like a lobster shell, security has layers — review code before you run it.
latestvk977s22500xactrvphfhe6bkc184tjmm
License
MIT-0
Free to use, modify, and redistribute. No attribution required.