llm-council — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited llm-council (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Run a generic multi-model council for any question. This is not coding-task specific.
node packages/lifeos-plugin/scripts/run-llm-council.mjs --prompt "<question>"Use --title "<short title>" when a clear artifact title is available.
To define a custom council, pass members explicitly. The number of council members is the number of --member values, or the first --council-size members when limiting a larger config.
node packages/lifeos-plugin/scripts/run-llm-council.mjs \
--prompt "<question>" \
--provider vercel \
--member "openai/gpt-5.4|GPT 5.4|openai" \
--member "moonshotai/kimi-k2.6|Kimi K2.6|vercel" \
--member "deepseek/deepseek-v4-pro|DeepSeek V4 Pro|vercel" \
--chairman "openai/gpt-5.4|GPT 5.4|openai"The runner prefixes shorthand model IDs with --provider, which defaults to vercel. For non-Vercel providers, either set --provider or include the provider as the third | field. Use openai/gpt-5.4 with provider openai for GPT 5.4 so it runs through the local OpenAI/Codex subscription, not Vercel.
runId.Configured in packages/lifeos-plugin/config/llm-council.normal.json:
openai/gpt-5.4vercel/moonshotai/kimi-k2.6vercel/deepseek/deepseek-v4-provercel/alibaba/qwen3.6-plusThe local environment must have:
opencode installed and authenticated for the configured providersLIFEOS_CONVEX_URL or CONVEX_URLLIFEOS_USER_IDLIFEOS_API_KEYUse --dry-run to inspect the prompt, model roster, and Convex save plan without calling models or writing artifacts.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.