AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Plan Feature Interview

skill-gtrabanco-agentic-workflow-plan-feature-interview · by gtrabanco

>

No reviews yet
0 installs
5 views
0.0% view→install

Install

$ agentstack add skill-gtrabanco-agentic-workflow-plan-feature-interview

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-gtrabanco-agentic-workflow-plan-feature-interview)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
1mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Plan Feature Interview? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Plan Feature — Interview (internal)

Turn a rough idea into a well-scoped, architecture-respecting feature through a focused interview. Optimizes for catching unknowns before implementation.

When to use

  • The plan-feature router calls this when the input is a vague idea (or the user

forces it with --interview) and wants help thinking it through before any code.

Step 0 — Discover the project (always first)

Per the agent guide's Workflow conventions + documentation map, then read what THIS skill needs: the architecture doc, the domain/style docs relevant to the idea's area, the SPEC template, and the roadmap. Goal: ask only what the project does not already answer, and ground every question in the project's real constraints (layering, domain/i18n/SEO/a11y rules, runtime/platform limits, naming conventions).

Process

  1. Restate understanding. Summarize the idea in one short paragraph — the

problem it solves and for whom. Get a yes/adjust before going deeper.

  1. Interview to fill the SPEC. Work through the dimensions below in small

batched rounds (use the question tool; group related questions; offer a recommended default each). Skip anything the docs already settle.

  • Problem & goal — what changes for the user; success signal.
  • Scope — explicitly what is OUT of scope.
  • Architecture impact — layers touched, new entities/ports, use-cases,

adapters; honor the dependency rules.

  • Data — schema/migrations, source of truth, cache/consistency.
  • Cross-cutting — i18n, SEO, a11y, domain rules, security, per the docs

map. Call out which apply.

  • UI reference (only when the feature has a UI surface) — is there a

design to build against (design-system components, mockup, Figma)? If the surface is significant and no reference exists, flag that a design pass should precede implementation rather than letting the code improvise.

  • Dev scenarios — happy path and failure modes (empty/degraded state,

races, outages) and how to reproduce each locally.

  • Acceptance criteria — objective, verifiable conditions for done.
  • Dependencies & risks — other features, external services, unknowns.
  • Non-goals / future work — deferred to issues, not built early.
  • Size — estimate XS / S / M / L (the SPEC template defines the scale).

XS/S → SPEC-only, single-pass execution; M/L → full artifact set, phased. If L, propose splitting into independently shippable features.

  1. Be proactive, not passive. Volunteer assumptions and risks the user

didn't mention; flag conflicts with existing architecture or docs; propose the smallest version that delivers the value.

  1. Traceability. Offer to open a tracking issue (from the feature issue

template); if created, the PR will Closes #n.

  1. Produce the SPEC. Once dimensions are answered, write the filled SPEC. The

plan-feature router then runs plan-feature-scaffold for the remaining artifacts + roadmap registration.

  1. Hand off — return exactly (fixed completion report, back to the router):

`` INTERVIEW → SPEC — size: Dimensions resolved: / Open questions: (→ decisions.md) | none Tracking issue: # | declined | n/a → scaffold next (plan-feature-scaffold) ``

Interview discipline

  • One topic per round; don't interrogate. 3–6 batched questions max, then

synthesize.

  • Always offer a recommended default so the user can move fast.
  • Don't ask what the SPEC template, architecture doc, or roadmap already answers.
  • Stop once the SPEC can be filled without guesses; capture residual unknowns as

open questions in decisions.md rather than blocking.

Guardrails

  • No code, no branch, no dependencies — design only.
  • Smallest-shippable mindset; defer extras to issues.
  • Respect and cite the project's architecture and style rules in your reasoning.
  • Otherwise per the project's Workflow conventions (docs-language).

Relationship to other skills

  • Sibling of plan-feature-from-issue (issue path); the plan-feature router

picks between them by input.

  • plan-feature-scaffold — scaffolds the artifacts once the interview yields a SPEC.
  • execute-phase — executes the phases afterward.

Done when

  • The interview resolved every SPEC dimension (or logged it as an open

question), a filled SPEC + artifacts exist, and the user knows the next step.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.