Shows sprint progress and task status. Use 'progress' (not 'status' - that's a built-in).
6 stars
0 votes
0 copies
1 view
Added February 10, 2026
developmentbashnodeapi
Works with
cli
api
Security analysis
A100/100
Scanned September 20, 2026
$npx -y skills add djnsty23/claude-auto-dev --skill status --agent claude-code
Installs into .claude/skills of the current project.
Are you the author of Status?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/djnsty23-status)
---
name: status
description: Shows sprint progress and task status. Use 'progress' (not 'status' - that's a built-in).
when_to_use: "Invoked when the user says \"progress\"."
allowed-tools: Bash, Read
model: haiku
user-invocable: true
---
# Status
Show current progress with minimal token usage.
## Sprint Data
Read the target project's PRD with the shared `workPlan(prd)` documented in
`core`. Report every sprint's records, not just the latest one or a file header.
Keep the current sprint and archived totals as separately labelled views.
Distinguish a missing file from unreadable/invalid data; neither is a clean queue.
For the "Blocked on you" line, run
`node "${CLAUDE_PLUGIN_ROOT}/scripts/prd-mark-needs-setup.js" --list` from the
project root: it reads every sprint and prints each needs-setup story with its
`blockedReason` and `blockedAt`.
## Process
1. Resolve the actual project and read the shared plan. If no PRD exists, report
that scope and any independently verified active work; do not invent a queue.
2. Show all five state counts plus unrecognised records and the total. Use
`plan.ready` for executable work and show `plan.blocked` / `plan.invalid`
reasons. Pending-state count and dependency-ready count are different.
3. Report active owners only from current ownership/worker evidence. A native
task list may supplement the report if this host actually exposes one, but
must not override the durable PRD or turn stale activity into a current fact.
```
[project] | [sprints covered] | [ref / working copy inspected]
Done: [N] | Pending: [N] | Failed: [N] | Deferred: [N]
Needs setup: [N] | Unrecognised: [N] | Total: [N]
Ready now: [N] | Dependency blocked: [N] | Invalid: [N]
Blocked on you: [N] ([ids]) — one line per story: id, what it waits for (blockedReason), since when (blockedAt)
Next: [id, title] | Active owner: [verified identity or unknown]
Unresolved: [ids and specific blockers]
```
## Proving the run
`done + pending + failed + deferred + needsSetup + unrecognised === total`
for the complete story population. Confirm the reported next story is present
in `plan.ready`. An empty population or no ready work with unresolved records
is not completion. Scores, native task-list emptiness and a green build do not
replace these checks. State the population and inaccessible data explicitly.
When `needsSetup` is nonzero, "Blocked on you" is printed as its own line with
ids, and its count equals `needsSetup`.
"Blocked on you" is a separate line from Pending on purpose. `[measured
2026-09-08]` six of the ten pending stories in one client repo were waiting on
a person — a pipeline variable, a partner's API, a decision — and had sat as
`passes: null` for up to 122 days, because the count that said "10 pending"
told nobody that an agent could advance only four of them. A needs-setup story
is remaining work for the operator and not for the agent; the line says who.