skills/intent/SKILL.md
Interactive conversation to clarify user intent step-by-step before creating specifications or solutions. Use when requirements are unclear or user has a new idea to explore.
npx skillsauth add sofer/.agents intentInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Guide the user through a structured conversation to deeply understand their intent before creating specifications or solutions.
Ensure alignment on:
Start by asking the user to describe their idea:
"Before we dive into specifics, could you describe your idea or the problem you're trying to solve in your own words?"
Listen actively and acknowledge their response before proceeding.
Ask each question one at a time, wait for response, then acknowledge:
Q1: Core problem/need
Q2: Primary users and expectations
Q3: Constraints and conditions
Summary checkpoint:
Q4: One-sentence goal
Q5: Measurable outcomes
When starting a new project, gather coding preferences:
Q6: Programming paradigm
Q7: Language and framework
Q8: Patterns and practices
Q9: Naming conventions
Summarise the complete intent:
Ask: "Does this fully reflect what you want to achieve?"
Once confirmed, create structured output for the manifest:
intent:
problem_statement: |
Clear articulation of the core problem being solved.
users:
- persona: "New hire"
needs:
- "Find onboarding materials quickly"
- "Complete setup without asking colleagues"
- persona: "HR manager"
needs:
- "Track onboarding progress"
- "Update materials easily"
constraints:
technical:
- "Must integrate with existing SSO"
- "Mobile-responsive required"
business:
- "Budget: internal team only"
- "Timeline: Q1 delivery"
organisational:
- "Must comply with data retention policy"
success_criteria:
- metric: "Onboarding completion time"
target: "< 2 hours"
- metric: "Support tickets about documentation"
target: "Zero in first month"
- metric: "User satisfaction score"
target: ">= 4.5/5"
coding_standards:
paradigm: "functional"
language: "TypeScript"
framework: "Next.js"
patterns:
- "Repository pattern for data access"
- "Dependency injection"
- "Event-driven for notifications"
forbidden:
- "No class inheritance deeper than 2 levels"
- "No mutable global state"
naming:
files: "kebab-case"
functions: "camelCase"
types: "PascalCase"
constants: "SCREAMING_SNAKE_CASE"
Save to .sdlc/manifest.yaml under the intent key.
Intent is a configured checkpoint. After completing:
tools
Check whether Claude and Codex have equivalent access to shared agent resources, skills, hooks, plugins, MCP servers, permissions, startup behaviour, and provider-specific adapter config. Use when comparing agent environments, debugging missing capabilities after restart, or deciding whether to symlink a resource or configure a runtime.
testing
Record substantive skill use in an append-only local log. Use after choosing or invoking a non-system skill for real work, when a skill is inspected but not used, or when a skill fails to apply. Do not use for routine system skills or incidental file reads.
testing
Turn a vague or underspecified request into a self-contained problem statement. Use when the user has a rough idea, when a request would fail if handed directly to an agent, or before non-trivial work that needs shared understanding.
data-ai
Append a one-line learning to ~/.agents/learning-log.md. Use when the user types /learning, or when something genuinely worth remembering surfaced during work and the user confirms it should be captured.