Groq API Inference
v1.0.0Build and debug Groq API chat and speech workflows with low-latency routing, structured outputs, and production-safe patterns.
⭐ 0· 471·3 current·3 all-time
byIván@ivangdavila
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 match the declared requirements: curl + jq and GROQ_API_KEY are appropriate for calling Groq inference endpoints and parsing responses. No unrelated services, binaries, or config paths are requested.
Instruction Scope
Runtime instructions are narrowly scoped to calling Groq endpoints, validating output, and storing small workflow files under ~/groq-api/. The docs explicitly advise not to store GROQ_API_KEY in files and to sanitize logs. This is coherent; the only minor note is that local logs and memory files could accidentally capture sensitive prompts if the user or agent chooses to persist them, so follow the guidance to sanitize before saving.
Install Mechanism
Instruction-only skill with no install spec and no downloads — lowest-risk installation model. It only relies on commonly-available tools (curl, jq) which are declared as required.
Credentials
Requests a single API key (GROQ_API_KEY), which is exactly what the skill needs. The SKILL.md does not reference other env vars or credentials beyond the declared one.
Persistence & Privilege
Does not request permanent 'always' inclusion, does not modify other skills, and only writes to its own directory under the user's home. Persistence is minimal and scoped to the skill's memory/log files.
Assessment
This skill appears coherent for Groq API work, but follow safe practices before installing: only provide a Groq API key (GROQ_API_KEY) and never paste it into saved files; restrict permissions on ~/groq-api/ (chmod 700) so local memory/log files are private; inspect any saved logs before sharing; confirm that the actual runtime agent you use will not autonomously exfiltrate data (the skill's docs say it limits network calls to Groq endpoints, but the agent environment enforces that); and avoid installing unrelated skills that request other credentials unless you need them. If you want extra assurance, run the example curl /models check yourself to verify expected behavior before enabling broader automation.Like a lobster shell, security has layers — review code before you run it.
latestvk977ehgfekhy3mjz3jk4nfedes81z628
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
⚡ Clawdis
OSLinux · macOS · Windows
Binscurl, jq
EnvGROQ_API_KEY