nozomio-labs / nia-skill

The scripts split across four kinds of work: indexing and grepping sources (repos.sh for GitHub, sources.sh as a unified front end for docs, datasets, local folders, Slack, and Drive, papers.sh for arXiv, datasets.sh for HuggingFace, packages.sh for npm/PyPI/crates.io/Go/RubyGems), search (search.sh for universal, per-source, and web queries), account and integration setup (auth.sh, slack.sh, google-drive.sh, folders.sh, categories.sh), and two features gated to a paid tier — oracle.sh, an autonomous research agent, and tracer.sh, live GitHub code search without indexing anything first. deps.sh, advisor.sh, contexts.sh, github.sh, and usage.sh round out the set with dependency tracking, a code advisor grounded in whatever has been indexed, cross-agent context sharing, direct GitHub API calls, and an API-usage summary.

Every script needs the same NIA_API_KEY, from a free signup, npx nia-wizard, or the auth.sh bootstrap commands, plus curl and jq on the machine running it — nothing here works without a Nia account. This is for someone who wants one agent-facing search layer across their own repos, their team's Slack, their Google Drive, and public docs and packages at once, rather than a narrower skill for each source separately.

READ THE FULL ANALYSIS＋

Two of the eighteen scripts sit behind a paid plan. The README marks oracle.sh and tracer.sh "(Pro)" with no further detail — no price, no free-tier limit, no description of what separates Pro from the base account beyond those two scripts. Everything else in the list reads as available on the same free signup that gets you an API key.

Checked 19 September 2026 against nia-skill's own README and its one skill description; we did not create a Nia account or run any script.