MiroFish Predict
Security checks across static analysis, malware telemetry, and agentic risk
Overview
The skill is coherent and purpose-aligned, but users should notice that it installs/runs external CLI and Docker components, uses an LLM API key, and may keep a local backend running.
This appears safe to use if you trust the MiroFish CLI and Docker image. Before installing, check the project/package source, use a limited LLM API key if possible, start with low simulation rounds to control cost, and stop the Docker backend when finished.
Static analysis
No static analysis findings were reported for this release.
VirusTotal
VirusTotal findings are pending for this skill version.
Risk analysis
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Installing or running the skill means trusting the MiroFish CLI package and its Docker image.
The skill relies on an external npm package and a Docker image pulled during first setup; these components are central to the stated purpose but are not included in the artifact contents.
"package": "mirofish-cli" ... "首次會自動拉 Docker image"
Install only if you trust the package/source, and consider reviewing the linked project and package provenance before use.
Simulations may use your LLM provider account and incur token costs.
The skill requires an LLM API key to run simulations, which is expected for its function but gives the CLI access to a billable credential.
需要 Docker Desktop 和 LLM API key ... `export LLM_API_KEY=xxx`
Use a scoped or low-limit API key where possible, monitor usage, and avoid sharing keys in prompts or reports.
A local backend may continue running until stopped, consuming local resources.
The skill documents a Docker backend that can be started and stopped; this is disclosed background behavior rather than hidden persistence.
`mirofish serve start` | 啟動 Docker 後端 ... `mirofish serve stop` | 停止後端
Use `mirofish serve status` and `mirofish serve stop` when finished.
