agentsclimarketplace

Audit user story

Skill abrahamFerga/scrum-skills/skills/audit-user-story

A vendor-agnostic library of AI agent skills built around the Scrum framework. Works with Claude Code, Cursor, GitHub Copilot, and any agentskills.io-compatible agent.

Install
npx -y skills add abrahamFerga/scrum-skills --skill audit-user-story

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Audits an existing user story against Scrum best practices — checks INVEST criteria, story format, acceptance criteria quality, and Definition of Ready compliance. Raises findings with clear severity levels and offers to post the audit as a comment on the work item for the Product Owner or PM to act on. Use when anyone says things like "audit this story", "review this user story", "is this story ready", "check this PBI", "does this story meet best practices", or pastes a story and asks for feedback. Don't use for creating new stories, sprint planning, or general Scrum questions. Works with ADO, Jira, or a pasted story.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

6.3 KB, as published. Nobody here has run it

User Story Audit

Purpose

A story that is unclear, too large, or untestable wastes the whole team's time — in refinement, in planning, and in delivery. This skill runs a structured audit so issues are caught before the story enters a Sprint, not during one.


Tool detection

  1. Check for active mcp__azure-devops__* tools → set $PM_TOOL to ado
  2. Otherwise check for active mcp__jira__* tools → set $PM_TOOL to jira
  3. If both → ask: "I see both ADO and Jira connected — which should I use?"
  4. If neither → set $PM_TOOL to manual

Step 1 — Fetch the story

  • ADO / Jira: Ask: "What's the work item ID or issue key you'd like me to audit?" Then retrieve the full item including title, description, acceptance criteria, state, and any linked items.
  • Manual: Ask the user to paste the story content directly.

Store the content as $STORY. If the item cannot be found, say so and ask the user to paste it instead.


Step 2 — Run the audit

Evaluate $STORY across four areas. Grade each check as:

  • Pass — meets the standard
  • ⚠️ Warning — present but weak; worth improving
  • Fail — missing or clearly broken; blocks readiness

Area 1 — Story Format

CheckPass condition
Title is action-orientedStarts with a verb + noun (e.g. "View monthly summary")
Story statement presentHas "As a / I want / So that" or equivalent
Persona is specificNot "user", "the system", or "admin"
"So that" states valueNames an outcome or benefit, not a restatement of the want
No implementation detail in the storyStory describes what, not how

Area 2 — Acceptance Criteria

CheckPass condition
ACs presentAt least 3 criteria exist
ACs are conditions, not UI stepsEach AC describes an observable outcome, not a sequence of clicks
ACs are independently verifiableEach AC can be tested without depending on another
Edge cases coveredError states, empty states, or boundary conditions addressed where relevant
No ambiguous languageAvoid "should", "might", "fast", "easy" — prefer measurable specifics

Area 3 — INVEST

CriterionWhat to look for
IndependentCan this be delivered without another unfinished story?
NegotiableIs the story a conversation starter, not a rigid spec?
ValuableDoes it deliver a real outcome to a user or the business?
EstimableEnough detail for developers to size it?
SmallCompletable within one Sprint by the team?
TestableCan the ACs be verified objectively?

Flag only the INVEST criteria that are genuinely at risk — not all six for every story.

Area 4 — Definition of Ready

CheckPass condition
Persona named and understoodTeam knows who this is for
Business value clearPO can explain why this is prioritized now
Dependencies identifiedBlockers or upstream dependencies are named
Size estimate presentStory has been sized (optional but flag absence)
No open assumptionsAny assumptions are captured and resolved

Step 3 — Present findings

Format the audit output as follows:

User Story Audit — [#ID / Issue Key] [Title]
Audited: [DATE]

Story Format
  ✅ / ⚠️ / ❌  [Check]: [brief finding]

Acceptance Criteria
  ✅ / ⚠️ / ❌  [Check]: [brief finding]

INVEST
  ✅ / ⚠️ / ❌  [Criterion]: [brief finding]

Definition of Ready
  ✅ / ⚠️ / ❌  [Check]: [brief finding]

Overall readiness: Ready / Needs work / Not ready

Top recommendations
1. [Most important fix — be specific, offer a rewrite if helpful]
2. [Second fix]
3. [Third fix]

Keep findings concise — a one-line note per check is enough unless a rewrite is warranted. Offer a specific rewrite for the persona, "So that" clause, or individual ACs when the existing text is weak.


Step 4 — Offer to comment

Ask: "Should I add these findings as a comment on the work item for the Product Owner to review?"

If yes, post a comment via the PM tool using wit_add_work_item_comment (ADO) or the equivalent Jira comment tool. Frame the comment as constructive peer feedback, not a rejection:

Story Audit Findings — [DATE]

This story was reviewed against Scrum best practices. Here are the findings:

[Paste the formatted findings block]

These are suggestions to help the team get the most value from this story — happy to discuss any of them in refinement.

If $PM_TOOL is manual, output the comment text in a clean block for the user to paste manually.


Guardrails

  • Never reject a story outright — frame every finding as an improvement opportunity.
  • Never suggest story points or estimates — sizing belongs to the Developers.
  • Never rewrite the full story without being asked — offer rewrites for specific weak sections only.
  • If the story is clearly an epic (covers multiple independent workflows), say so clearly: "This reads like an epic — the most valuable single story might be [X]. Want me to help split it?"
  • Keep the tone peer-to-peer. The audit is a team quality tool, not a gate kept by a single role.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.