agentleFS
edilson-silvaedilson-silva/codride/.agents/skills/product-spec/SKILL.md

product-spec

Skill2 starsChanged 2 months agoMITView on GitHub

Expand a validated requirement into a full PRD

We split this skill into the rules it gives an agent

4 rules. 0 are common, 2 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 9 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If the initial requirements aren't enough to move on to a PRD, ask the user clarifying questions and update the requirement artifact/issue …

    line 19
  2. Very specific:

    Keep scenarios concrete and testable — they become the contract engineer-work implements against and engineer-coverage checks against.

    line 33
  3. Distinctive:

    Don't overthink this. If the issue doesn't call for some of these, skip them — less is more, but don't drop important details.

    line 40
  4. Very specific:

    If docs/business-context/features/ doesn't exist yet, create it.

    line 72

“Common” is measured: kept by at least 6 in 10 of the 9 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

You're acting as a product specialist helping a human take an issue's requirement to its final stage, ready for build.

If this doesn't already have a GitHub issue and refined requirement behind it, and the task is small/well-understood, product-quick-spec produces the same output (FR-XX + BDD feature file) in one step — consider redirecting there instead of forcing this staged process on something that doesn't need it.

Your goal is to take an initial requirement and understand it deeply.

The user's request should point to the requirement to expand — a GitHub issue number, or the requirement text itself. If neither is given, ask which one they mean rather than guessing.

  1. Validate the current requirements. Review what you were given and confirm it contains the essentials:

    • WHY we're doing this
    • WHAT is being built
    • HOW it's being built — less important than the other two, but good to have a sense of it

    If the initial requirements aren't enough to move on to a PRD, ask the user clarifying questions and update the requirement artifact/issue before proceeding. Don't assume anything — ask.

  2. Check the project's master docs to identify any specific rule to follow, or whether this request violates a master doc. If it does, ask for clarification and only proceed if the user says to.
  3. Build your understanding of the key PRD elements:

    • Product overview: problem statement and market opportunity, target users and personas, product vision and goals, success metrics and KPIs
    • Functional requirements: core features and capabilities, user stories or use cases, user flows and interactions, technical specs, API requirements (if applicable). Number each discrete requirement FR-01, FR-02, ... — this numbering carries through to the acceptance criteria below and to the feature file in step 6.
    • Acceptance criteria (BDD): for each FR-XX, write the scenarios that define "done" as Given/When/Then:
      Given [starting state]
      When [the user does something]
      Then [expected outcome]
      Keep scenarios concrete and testable — they become the contract engineer-work implements against and engineer-coverage checks against. Cover the happy path plus the edge cases that actually matter (don't pad with trivial ones). Skip this for requirements too small to need it.
    • Non-functional requirements: performance benchmarks, security/compliance needs, scalability requirements, accessibility standards
    • Design and UX: UI/UX guidelines, wireframes or mockups, design-system references, platform-specific considerations
    • Technical considerations: architecture overview, integration requirements, data requirements, third-party dependencies
    • Project details: risks and mitigation strategies, release criteria and rollout plan
    • Constraints and assumptions: technical constraints, business constraints, key assumptions being made

    Don't overthink this. If the issue doesn't call for some of these, skip them — less is more, but don't drop important details.

  4. Present your understanding to the user, along with any clarifications needed. Iterate until you have 100% clarity.
  5. Once the user approves, edit the requirement artifact/issue/file, enriching it with what was discovered.
  6. Save (or update) docs/business-context/features/<slug>.md — this is what makes the feature visible to product-sync-github; without it, the PRD only lives on the issue and never gets tracked there. Use a kebab-case slug derived from the feature title. Format:

    # [Feature Title]
    
    **Status**: Planned
    **Priority**: [Critical | High | Medium | Low]
    **Scope**: [MVP | Phase 2 | Phase 3 | ...]
    **Module**: [if this project groups features by module; omit otherwise]
    
    [1-2 paragraph product overview]
    
    ### FR-01: [Requirement title]
    [Description]
    
    **Acceptance criteria:**
    ```
    Given ...
    When ...
    Then ...
    ```
    
    ### FR-02: [Next requirement]
    [Same structure]

    If docs/business-context/features/ doesn't exist yet, create it. If the file already exists (this feature was specced before), update it rather than duplicating.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.