---
name: source-command-prd-track
description: Capture the current work as / into a TRK-* tracking record now (mid-session). Host-native /prd-track equivalent for hosts that load skills instead of Claude command files; trigger on "/prd-track", "prd-track", or a request to run the prd-track operation.
---

# /prd-track (host-native command skill)

Generated from `commands/prd-track.md` by `scripts/generate_command_skills.py` — do not edit by hand; edit the command file and regenerate.

Treat the user's invocation as this command. Where the instructions use `$ARGUMENTS`, substitute the arguments the user supplied.

Update tracking for the work in progress, without waiting for session close.

Use the `project-memory` skill to create or update a `TRK-*` record in
`.prd_plugin/state/tracking.json` for the current work: a concise summary,
`type` (`goal` for the session's overall objective, `active_work` for a sub-task),
status (`active`/`complete`), and `linked_ids` to the active `REQ-*` /
`IMP-TASK-*` / `CHG-*`. Set `type: goal` and `status: active` when capturing the
thing you are driving to completion — that open goal is what the autonomous Stop
guard keeps working toward. If a matching active `TRK-*` already exists, update it
instead of creating a duplicate. Register any new ID in
`.prd_plugin/ids/registry.json`. Use complexity/confidence ratings, not time
estimates. Then report the TRK id and what you recorded.

Optional focus from the user: $ARGUMENTS

## Staleness

This adapter owns no records of its own; apply the shared policy in `.prd_plugin/method/staleness-rules.md` to any records the underlying operation reads or writes.