Skip to content

feat(commands): add /prd command for requirement crystallization#22

Open
JYHTHEGREAT wants to merge 1 commit into
sangrokjung:mainfrom
JYHTHEGREAT:feat/prd-command
Open

feat(commands): add /prd command for requirement crystallization#22
JYHTHEGREAT wants to merge 1 commit into
sangrokjung:mainfrom
JYHTHEGREAT:feat/prd-command

Conversation

@JYHTHEGREAT

Copy link
Copy Markdown

Context

Contribution from jyh-system, a production fork running 6+ projects over ~1 month. This addresses a gap in the forge workflow: there's no structured step between a vague idea and /plan.

What This Adds

/prd command — crystallizes vague user ideas into structured requirements before planning.

Key Features

  • 5-axis probing: Problem, Expected Outcome, Boundaries, Constraints, Priority
  • 3 depth levels: Lite (inline), Standard (file), Full (scenarios + non-functional)
  • Readiness Gate: All acceptance criteria must be testable/observable before proceeding
  • Scope Guard: Auto-detects multi-subsystem requests and suggests decomposition
  • Anti-patterns: One question at a time, no guess-filling, no premature implementation hints

Why This Matters

The second most common AI coding failure mode is "start implementing from vague requirements → wrong direction → rework". /prd structurally prevents this by forcing intent clarification before /plan.

Changes

  • commands/prd.md (193 lines) — new file, no existing files modified

Existing Code Impact

None. Pure addition. No existing commands, agents, or workflows are modified.

Test Plan

  • Run /prd "add user authentication" → should ask structured questions, not jump to implementation
  • Run /prd --depth lite "fix typo in README" → should produce inline 3-item PRD
  • Verify Readiness Gate blocks vague acceptance criteria

🤖 Generated with Claude Code

Adds a structured workflow to crystallize vague ideas into clear
requirements before planning. Includes 5-axis probing (Problem,
Outcome, Boundaries, Constraints, Priority), 3 depth levels
(Lite/Standard/Full), and a Readiness Gate that prevents
proceeding with ambiguous requirements.

This addresses the common failure mode where implementation starts
from unclear requirements, leading to rework cycles.

Based on production usage in jyh-system (~1 month, 6+ projects).
sangrokjung added a commit that referenced this pull request May 3, 2026
…lize)

PRs opened in parallel via 3 background workers (1a/1b/1c plus
SendMessage continuation for composiohq option B):

- travisvn:  travisvn/awesome-claude-skills#679
- rohitg00:  rohitg00/awesome-claude-code-toolkit#362
- composiohq: composio-community/awesome-claude-plugins#210
  (replaces dormant PR #22 from 2026-02-28; #22 closed with backfill)

Deferred (intentional):
- hesreallyhim — HOLD pending upstream TOC stabilization (currently 'TODO' stub)
- voltagent / ccplugins — Strategy A 30-min in-repo contributions (next sprint)

Acceptance criterion #2 (≥3 PRs opened) MET. Plan COMPLETED.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant