From enterprise-harness-engineering
Generates professional User Stories via guided interviews to gather background, goals, and testable acceptance criteria; places files in docs/04-user-stories/ directories.
npx claudepluginhub addxai/enterprise-harness-engineering --plugin enterprise-harness-engineeringThis skill uses the workspace's default tool permissions.
This skill embodies the "Everything is Code" engineering philosophy, helping you produce high-quality user stories and their supporting documentation alongside your MR submissions.
Generates user stories with Given/When/Then acceptance criteria following INVEST principles from PRDs, feature-persona-goal inputs, narratives, Jira epics, or screenshots.
Writes user stories from user perspective with acceptance criteria. Use for Agile backlog creation, sprint planning, requirement breakdown, and defining done.
Provides user story templates using As-a/I-want/So-that format, Given-When-Then acceptance criteria, story splitting, and INVEST criteria for agile backlog refinement and requirements definition.
Share bugs, ideas, or general feedback.
This skill embodies the "Everything is Code" engineering philosophy, helping you produce high-quality user stories and their supporting documentation alongside your MR submissions.
For the detailed example template, see examples/TEMPLATE-UserStory.md.
git status to determine the relevant module and place documents in the corresponding docs/04-user-stories/ directoryUser requests "help me write a User Story," "describe this feature," "generate user story," etc.
After receiving a request, check whether the following three elements are present:
| Element | Meaning | Sufficiency Criteria |
|---|---|---|
| Background (Why) | Current pain points, impact scope | Can articulate "who is affected in what scenario" |
| Goals (What) | Verifiable success metrics | Can articulate "what can be quantified after completion" |
| Acceptance Criteria (AC) | Key boundary scenarios | At least 2 checkable conditions are listed |
If information is insufficient, execute Step 2 (interview); if sufficient, proceed directly to Step 3 (generation).
Take on the dual role of "detective" and "poet," asking all missing dimension questions in a single round (avoiding multi-round fragmented questions):
To generate a professional user story, I need you to confirm the following (keep it brief, 1-3 sentences per item):
1. **Background/Pain point**: [specific missing background question]
2. **Goal**: [specific missing goal question]
3. **Acceptance criteria**: [specific missing AC question]
You can describe briefly — I will handle converting it into the standard professional document format.
Strictly follow the structure from TEMPLATE-UserStory.md to generate the file.
File naming convention: {feature-name}.md (lowercase kebab-case)
Placement path: {submodule_root}/docs/04-user-stories/{feature-name}.md
If the target directory does not exist, first execute Feature B or prompt the user.
User explicitly requests "create docs structure," "initialize docs," etc.
docs/
├── 01-product-research/ # Product research and competitive analysis
├── 02-system-design/ # System design and architecture decisions
├── 03-detailed-design/ # Detailed design documents
├── 04-user-stories/ # User Stories
└── 05-user-guide/ # User manuals and operation guides
.gitkeep file in each subdirectory to ensure Git tracks empty directoriesAGENTS.md in the docs/ root directory describing each directory's purposedocs/AGENTS.md Content Template# docs/
This directory is the project documentation root, organized by R&D lifecycle stages.
## Directory Structure
- `01-product-research/`: Product research, competitive analysis, market insights
- `02-system-design/`: System architecture design, technology selection, ADR (Architecture Decision Records)
- `03-detailed-design/`: Module detailed design, interface specifications, data structures
- `04-user-stories/`: User Story files, naming convention: `{feature-name}.md`
- `05-user-guide/`: End-user operation manuals and usage instructions
## Maintenance Conventions
- User stories are created as individual `.md` files per feature module
- This file should be updated after architectural/directory-level changes
Analyze code changes or user descriptions to determine the target sub-module in the following order:
device-cloud-server/, frontend/)git status output## User Story: Alarm Filtering Feature
**As** a system administrator, **I want** to filter alarms, **so that** I can reduce distractions.
**Acceptance Criteria**:
- [ ] System supports filtering functionality
- [ ] Filtered alarms are not displayed after filtering
Problem: Missing background, ACs are too vague, unable to verify "what conditions trigger filtering."
## User Story: Rule-Based Alarm Silence Filtering
**Background**: The monitoring system generates 500+ heartbeat probe alarms daily, causing
operator alert fatigue. Real fault alarms get buried, and average response time has increased
from 2 minutes to 15 minutes.
**As** an operations engineer, **I want** to configure alarm silence rules (by device type + alarm type combination),
**so that** known non-fault alarms are suppressed and real fault alarm response time returns to under 2 minutes.
**Acceptance Criteria**:
- [ ] Given a configured silence rule "device type=heartbeat probe AND alarm type=connection timeout,"
When such an alarm fires, Then the Grafana dashboard does not display it and the audit log records "silenced"
- [ ] Given the silence rule configuration interface, When the admin saves a rule,
Then the rule takes effect within 30 seconds without requiring a service restart
- [ ] Given an alarm matches both a silence rule and an escalation rule simultaneously,
When the alarm fires, Then the silence rule takes priority, but the conflict is noted in the audit log
| Scenario | Condition |
|---|---|
| Technical refactoring MR | No visible user-facing functionality change — may simplify to a brief technical note |
| Pure bug fix | Reference the corresponding Issue directly — full User Story format not required |
Exemption method: /override skill=story-craftsman reason="pure technical refactoring, no user-visible functionality change"