Precision Coral Metrics API

Analyzes underwater images using YOLOv11 and MobileSAM to precisely segment coral colonies and calculate accurate coral coverage metrics.

Security Scan

The skill is internally consistent: it delegates heavy ML work to an external OpenClaw gateway, asks for no local secrets or installs, and the provided example matches that design (minor documentation mismatches noted).

Purpose & Capability

The skill describes GPU-heavy models (YOLOv11 + MobileSAM) but does not include model code — instead it calls a remote OpenClaw API gateway, which is coherent for a hosted inference service. Nothing in the package requests unrelated credentials or system access. Minor documentation inconsistencies: the SKILL.md integration example shows POST to /v1/skills/coral-ai/predict while the usage_example.py and other documentation reference /v1/skills/coral-precise-coverage-ai/predict; SKILL.md header version (1.1.0) differs from registry version (1.0.1). These look like copy/edit issues, not malicious misdirection.

Instruction Scope

Runtime instructions and the example script only instruct the agent/user to upload an image to the OpenClaw gateway with an X-OpenClaw-Token and to handle the JSON/base64 result. The instructions do not ask the agent to read system files, environment variables, or contact other endpoints. The only scope concerns are the aforementioned endpoint/path and version wording inconsistencies in documentation.

Install Mechanism

No install spec (instruction-only) and a single example script — nothing is downloaded or written to disk by an installer. This is the lowest-risk install posture.

Credentials

The skill declares no required environment variables and no primary credential in registry metadata, yet SKILL.md and usage_example expect an API token supplied as X-OpenClaw-Token (passed at call time). This is proportionate for a gateway-based service; the only minor inconsistency is that the registry metadata doesn't list the token as a required/primary credential (documentation omission).

Persistence & Privilege

always:false (default) and no code attempts to persist or modify other skill/system settings. The skill does not request persistent presence or elevated privileges.

Assessment

This skill appears to be a hosted inference service that sends your images to the OpenClaw gateway for processing — it does not run models locally or request system credentials. Before installing or using: (1) Confirm you trust the OpenClaw gateway URL (https://api.openclaw.io) and the skill developer for handling potentially sensitive images. (2) Verify billing/pricing (pay-per-invocation) and storage/retention policy for uploaded images and results. (3) Note minor documentation mismatches (endpoint path and version) — confirm the correct endpoint with the provider before automating uploads. (4) Ensure you supply and store your X-OpenClaw-Token securely (the skill itself does not require other secrets). If you need offline/local processing or guaranteed zero-exfiltration, this hosted skill is not appropriate.

Precision Coral Metrics AI (CM-AI)

⚡ Overview

CM-AI is an industrial-grade vision skill leveraging the YOLOv11 + MobileSAM hybrid architecture. Designed for marine ecologists and environmental agencies, it replaces error-prone manual reef assessments with rapid, pixel-perfect coral coverage analysis.

🚀 Key Capabilities

1. High-Fidelity Detection: Precisely locates coral colonies in complex, high-noise underwater backgrounds using YOLOv11.
2. Pixel-Perfect Segmentation: Leverages MobileSAM for refined mask extraction, ensuring accurate area calculation even for overlapping organisms.
3. Automated Metrics: Instantly calculates the Coral Coverage Ratio (%) and identifies individual colony counts.

📈 Roadmap

• v1.2.0: Genus-level identification (e.g., Acropora, Brain Coral, Montipora).
• v1.3.0: Fully automated transect data extraction and biodiversity index analysis.

🔒 Security & Access (OpenClaw Protected)

[!CAUTION] This skill is strictly integrated with the OpenClaw API Gateway. To protect backend GPU resource integrity, direct non-gateway traffic will be automatically blocked. Developer royalties are settled via the OpenClaw monetisation protocol.

🛠️ Integration Example

All requests must follow the OpenClaw standard authentication:

POST https://api.openclaw.io/v1/skills/coral-ai/predict
Headers: X-OpenClaw-Token: <YOUR_TOKEN>
Body: multipart/form-data (key: 'file')

---

© 2026 @mingzhuangwang | Powered by OpenClaw Ecosystem