Ops

Ops Status Report

Compiles the status report your folder-based workspace already contains but nobody assembles: active items, overdue follow-ups, recent activity, tool health. Read-only except the report file, ending in an Attention Needed section that is the whole point.

/ops-status-report Free
X-Ray: what this skill can and cannot do
Shell access Yes (read-only commands you configure, such as a CLI status check)
Network calls No
File writes The dated report file, and the config file once (with your approval)
File reads Folder listings, markdown frontmatter, and file dates in the configured sections
Destructive ops No; the report never fixes what it finds, it names it
Credential access No (a command that needs a missing .env is reported as not run, and the report continues)
Scope One configured workspace per run

Why this exists

The first status report I ran on my own workspace took two minutes and surfaced eleven overdue outreach follow-ups, most of my client folders missing a brief file, and a lead pipeline that had quietly gone empty. None of it was hidden. Every one of those facts was sitting in a folder or a frontmatter key I could have read any day. Nobody had compiled it, so functionally nobody knew it.

That is the gap this skill fills. A workspace full of markdown and folders holds an honest status report at all times; it just never gets assembled. This skill assembles it on demand: what is active, what is overdue, what went quiet, whether the tools still run. The Attention Needed section at the end is the whole point, and it has one rule I care about more than any other: if everything is genuinely fine, it says so in one line instead of inventing urgency. A report that manufactures alarm gets ignored within three weeks, and then it misses the real one.

How it works

Load the Config
Collect Read-Only
Write the Report
Attention Needed
  1. It learns what "status" means to you, once On first run it interviews you: which folders are tracked items, which frontmatter keys hold status and due dates, which folders should show recent activity, which read-only commands prove a tool is healthy. Then it offers to save that as a config so every later run is instant.
  2. It collects without touching anything Every configured section runs read-only. Missing files are findings ("this item has no status file"), not errors to skip. Unparseable dates are listed as-is, never guessed. A failing command gets one line with the reason, and the report continues.
  3. It writes a dated report Tables for tallies, lists for activity, short summaries for command health. Saved to your reports folder with the date in the filename, so the reports themselves become a timeline.
  4. It ends with Attention Needed Three to five bullets maximum, ordered by the cost of ignoring them: overdue items first, then untracked items, then things that went quiet. If everything is fine, it is one line saying so.

Step by step (for first-time users)

  1. Open Claude Code in your workspace and type /ops-status-report Client folders, a lead pipeline, project trackers: any folder-based workspace where status lives in files.
  2. Answer the one-time interview What counts as a tracked item, where its status lives, what "overdue" means, and which commands to health-check. Five minutes, once.
  3. Approve the saved config It writes your answers to a small config file in the project root, so next week the report runs with no questions at all.
  4. Read the Attention Needed section first It is shown directly in chat with the report path. The rest of the report is supporting evidence for those bullets.
  5. Act on it yourself, then rerun weekly The report names problems, it never fixes them. Put it on your scheduler of choice and the compiled truth arrives every week.

What a section can check

Section typeWhat it answers
Folder scanWhich tracked items exist, and what status each one declares
Frontmatter tallyCounts by status key, plus which date key marks an item overdue
Recent activityWhat was created or modified in the last N days, or went quiet
CommandRead-only output of a tool's own status check, with failures explained
Reports folderWhere the dated reports accumulate into a history

Honest take

What it does well: The ordering discipline in Attention Needed. Three to five bullets, ranked by the cost of ignoring them, is a genuinely different artifact from a dashboard of everything. And the honesty rules compound: numbers come from counting, never estimating; a folder that could not be read says it could not be read; a fine week is reported as a fine week. That restraint is what keeps the report trusted enough to still be read in month three, which is when it catches the thing that matters.

What it does not do: It does not fix anything it finds, by design; a report that quietly repairs things stops being a report and starts being a liability. It does not reach into email, calendars, or any external service: if the status is not in your files or a local command's output, it is not in the report. And it is only as good as your file discipline. A workspace where nothing carries a status or a date will produce a report that mostly says "untracked".

When to use it: Weekly, on a schedule, in any workspace where work lives in folders: freelance client rosters, lead pipelines, content calendars, multi-project setups. And immediately after any long heads-down stretch on a single project, because that is exactly when the rest of the workspace has been quietly accumulating overdue items.

Use this skill in your project

Download the .md file, drop it into .claude/skills/, and run /ops-status-report.

Download .md
Share
X LinkedIn Reddit