AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Triage

skill-aldianriski-lean-flow-triage · by aldianriski

Use when the existing backlog needs grooming — re-prioritising tasks, flagging stale/duplicate/conflicting items, marking what is ready vs blocked vs needs-info, and routing rejected work out of scope. Operates on TODO.md Backlog only, never the Active Sprint. Do not use to create new tasks (use /task-decomposer) or to form a sprint from the groomed backlog (use /lean-doc-generator promote).

No reviews yet
0 installs
10 views
0.0% view→install

Install

$ agentstack add skill-aldianriski-lean-flow-triage

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-aldianriski-lean-flow-triage)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
1mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Triage? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

triage

Groom the existing backlog: re-rank, flag, state, and route. Backlog only — never re-orders a locked Active Sprint. Re-prioritisation and state changes are HITL: propose with reasons, get a y, then apply.

Pipeline: /task-decomposer (intake) → /triage (groom + re-prioritise) → /lean-doc-generator promote (form sprint) → /orchestrator (build).

When to invoke

  • The backlog has grown and priorities have drifted from reality.
  • Before a Sprint Promote, to surface what is genuinely ready.
  • A specific task needs a state or priority decision.

Task states (light)

A Backlog task carries an optional state: — orthogonal to the HITL/AFK label (which says who acts, not whether it's ready):

| state | meaning | |---|---| | ready | fully specified, has a done-when — promotable | | needs-info | missing detail or an unanswered question — not promotable yet | | blocked | waiting on a dependency or another task (depends-on) |

Rejected work is not a state — it leaves the backlog (see .out-of-scope/). Default when unset: ready if it has a done-when, else needs-info.

Flow

  1. Scan .out-of-scope/ first — if a backlog task resembles a prior rejection, surface it and ask before keeping it.
  2. Load — read TODO.md § Backlog (+ § Tech Debt). Ignore the Active Sprint.
  3. Re-rank — re-evaluate each task's P0–P3 tier; propose moves with a one-line reason each. Order within a tier by dependency, then impact × urgency.
  4. Flag — surface stale (no movement / superseded), duplicate (same concern → merge or differentiate), conflict (acceptance criteria that contradict another task).
  5. State — set ready / needs-info / blocked; for needs-info list the specific questions; for blocked name the blocker.
  6. Route rejects — for work that will not be done, write an .out-of-scope/ entry, then remove the task from the backlog with a pointer.
  7. Output — the groomed backlog + a ready-to-promote shortlist (P-order, ready only). Apply changes only after human y.

.out-of-scope/ knowledge base

Rejected enhancements are remembered, not forgotten — so they are not re-litigated next quarter. Create lazily (only on the first rejection). **One file per concept, not per task** — repeat requests for the same thing accumulate under one file's prior-requests list.

.out-of-scope/.md
# 
- date: YYYY-MM-DD
- decision: out of scope
- reason: 
- revisit-if: 
- prior-requests: TASK-012, TASK-031, …
  • Durable reason, not a deferral — "we're too busy right now" is a deferral, not a rejection; don't file it. The reason must still hold next year.
  • Match by concept, not keyword — "night theme" matches dark-mode.md. On a match, surface it: "resembles .out-of-scope/.md, rejected because … — still true?"confirm (append to prior-requests, drop the task) · reconsider (delete the file, task proceeds) · disagree (related but distinct, task proceeds).

Hard rules

  • Backlog only — never re-order or re-state the Active Sprint (a sprint in flight is locked).
  • HITL — propose re-ranks, state changes, merges, and rejections; apply only after y.
  • Never silently delete — a removed task goes to .out-of-scope/ with a pointer, or is merged with a note. No task vanishes without a trail.
  • Don't re-decompose — triage grooms existing tasks; new tasks come from /task-decomposer.

Red flags

Re-prioritising the Active Sprint — a sprint is locked; grooming is a Backlog activity. ❌ Deleting a task with no trail — route to .out-of-scope/ or merge with a note. ❌ Promoting a needs-info / blocked task — only ready work is promotable. ❌ Re-ranking without a reason — every priority move carries a one-line justification. ❌ Ignoring .out-of-scope/ — re-litigates settled rejections; always scan it first.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.