Feature planning
Skill ahtishamshahzad/agent_dev_flow/.ai/skills/feature-planning
Use to plan a feature work item — scoping new capability within existing (or planned) architecture, defining its phases, tasks, tests, and acceptance criteria without expanding scope. Delegates general phase/task mechanics to task-planning.From its SKILL.md
npx -y skills add ahtishamshahzad/agent_dev_flow --skill feature-planningAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 26 days oldThe repository was created 26 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 0 stars0 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.
SKILL.md
3.6 KB, 718 tokens by cl100k_base, as published. Nobody here has run it
Feature Planning
Purpose
Plan a feature: define its scope within the architecture, its design decisions, tasks, and tests, so it can be implemented and verified without scope creep. Produces a feature work item in ../../work-items/features/.
When to Use
- Request classified as feature (or a feature sub-part of a larger plan).
- When adding capability to an existing or planned architecture.
- Not for defect fixes (
bug-investigation) or behavior-preserving cleanup (refactor-planning).
Inputs
- Requirement baseline + success criteria.
- Architecture + repository layout.
- Existing patterns (from
existing-project-audit) for consistency.
Discovery Questions
- What exactly is in scope, and what is explicitly out?
- Which applications/modules does it touch?
- What are the acceptance criteria from the user's perspective?
- What tests prove it works (levels/tools)?
- Any security-sensitive surface (auth, payments, PII)?
Responsibilities
- Define the feature's scope and boundaries (in/out).
- Fit it to existing architecture and patterns.
- Produce tasks (via
task-planning) with acceptance criteria. - Attach a testing plan (
testing-strategy) and security requirements (security-review) where relevant. - Record the feature work item and its status.
Required Workflow
- Read requirements + architecture.
- Lock scope (in/out) and acceptance criteria.
- Identify affected apps/modules and contracts.
- Generate tasks (delegate mechanics to
task-planning). - Attach tests + security requirements.
- Record in
../../work-items/features/.
Decision Rules
- Match existing conventions rather than introducing new patterns without cause.
- Security-sensitive features get integration/E2E tests, not just units.
- New work discovered mid-plan → propose a separate work item, don't absorb it.
Rules
- Stay within the feature's stated scope.
- Preserve unrelated behavior.
- Acceptance criteria are observable and testable.
Anti-Patterns
- Scope creep ("while I'm here…").
- Introducing a new stack/pattern for one feature.
- Shipping without acceptance criteria or tests.
Validation Checklist
- Scope (in/out) explicit.
- Affected apps/modules identified.
- Acceptance criteria observable/testable.
- Tasks generated with criteria.
- Testing plan attached.
- Security requirements attached where relevant.
Definition of Done
A recorded feature work item with locked scope, acceptance criteria, tasks, a testing plan, and any security requirements — ready for implementation under the gates.
Related Skills
task-planning, architecture-design, testing-strategy, security-review, code-review, project-orchestrator.
Related Knowledge
../../knowledge/ (domain rules, patterns).
Related References
../../references/<domain-topic>/ when the feature domain needs grounding.
Context Loading Guidance
- Requires: requirements, architecture, affected-area patterns.
- Does not require: unrelated modules, the full reference tree, other planning skills' bodies.
- May load:
task-planning,testing-strategy,security-review. - Stop when: the feature work item is recorded.
Token Efficiency Guidance
Load only the affected modules' context. Reuse task-planning for mechanics rather than restating them. Keep the work item concise.
What ships with it
Read from the repository
Just SKILL.md. No reference files, no scripts.