Coala Client
PassAudited by ClawScan on May 1, 2026.
Overview
This is a coherent instruction-only guide for using the Coala CLI, with expected cautions around installing an external CLI, using API keys, and importing MCP tools or skills.
This skill looks safe to install as documentation for the Coala CLI. Before using it, make sure you trust the coala-client package, only import MCP toolsets or skills from trusted sources, and handle any OpenAI/Gemini API keys carefully.
Findings (4)
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 the skill’s required CLI gives that package local execution capability when the user runs it.
The skill depends on installing an external CLI package. This is central to the stated purpose, but users should trust the package source and version they install.
uv | package: coala-client | creates binaries: coala-client
Install from the official package source, verify the package name and version, and keep it updated.
The CLI can use the user’s LLM provider account and may incur provider usage or expose prompts to that provider according to the provider’s terms.
The skill documents provider API key use for LLM chat. This is expected for OpenAI/Gemini integration, and the artifacts do not show hardcoding, logging, or unrelated use.
Set API key e.g. `export OPENAI_API_KEY=...` or `export GEMINI_API_KEY=...`. Ollama needs no key.
Use scoped or revocable API keys where possible, avoid sharing config files containing keys, and revoke keys if they are exposed.
If a user imports an untrusted skill, its instructions could shape later chats when loaded.
Imported skills are stored persistently and can later be loaded into the chat context. This is a core feature, but untrusted skill text could influence future agent behavior.
`coala skill <SOURCES...>` ... `SOURCES: GitHub tree URL ... zip URL, or local zip/dir.` ... `/skill <name>` — load skill ... into context.
Import skills only from trusted sources and review their SKILL.md content before loading them into a chat.
Prompts or tool arguments may be sent to configured MCP servers/tools during use.
The skill documents interaction with MCP servers and tools. This is purpose-aligned, but MCP tools can receive user-supplied data and should be trusted.
`coala` or `coala chat` — interactive chat with MCP tools. ... `coala mcp-call <SERVER>.<TOOL> --args '<JSON>'`
Use MCP servers from trusted sources and review configured servers before sending sensitive data through them.