agent-narrator

Watch the agent work: every step in plain English, with a live time-saved counter.

Translates every tool call into one plain-English sentence ("Reading the pricing page to find the old numbers") and keeps a running estimate of time saved. /narrate smart uses a model call per step; the default is rule-based and free. /narrate demo plays a scripted session. We built this for training sessions where the audience has never seen an agent work.

Install

In a Claude Code session (2.1.287 or later):

/plugin marketplace add OneWave-AI/claude-code-mods
/plugin install agent-narrator@claude-code-mods

Or load this folder for one session: claude --plugin-dir ./agent-narrator

What it reaches

| Mod | Network | Runs processes | Files | Calls a model | Sends data anywhere | | --- | --- | --- | --- | --- | --- | | agent-narrator | No | No | No | Only with /narrate smart: the tool name and short fields (path, command), never file contents | Only to your Claude model |

Run claude plugin validate ./agent-narrator to list every event it hooks and every call it makes. See the repository README for the full security notes.

Part of claude-code-mods by OneWave AI. MIT licensed.