This skill wraps `jc` as a bounded agent workflow: take raw output from established command-line tools or supported file types, convert it into structured JSON, then pass that JSON to later automation steps. An agent should invoke it when the source of truth already exists as human-readable shell output, but the next step needs reliable structure for filtering, comparisons, policy checks, or storage. That is the sweet spot for inventory snapshots, incident triage, compliance evidence collection, and pre/post change diffs.

The important scope boundary is that this is not a general shell framework, remote-execution layer, or replacement for the original command. The agent is not inventing a new data source. It is using the upstream tool `jc` to normalize existing command output or supported files into JSON so downstream code stops depending on fragile grep, awk, or regex chains. If the workflow already has a native API or a command that can emit JSON directly, the agent should prefer that instead.

Integration is simple and practical. Run the original command locally, over SSH, or inside CI, pipe the result to `jc`, then hand the JSON to `jq`, Python, Ansible, a database loader, or a reporting step. The same pattern works for fleet audits, host diagnostics, and agent runbooks where one step gathers evidence and later steps summarize or enforce policy. Upstream sources confirm `jc` is maintained by Kelly Brazil, published on PyPI, released on GitHub, and installable via `pip3 install jc`, OS packages, or release binaries.