prd-full

Savant verdict: Review advised

Security analysis or source-hub review identified patterns requiring attention.

Live evaluation

Not evaluated yet. Workspaces can request a live evaluation.

Security assessment

Scan pending.

Structure

• No license declaredConfirm you may reuse this skill before importing it into your repository.
• The source hub's security checks flagged this skillGen Agent Trust Hub: warn (The skill contains instructions that explicitly override built-in safety protocols and human confirmation prompts. While intended for automation, this removes critical oversight during the execution of external tasks, which could allow malicious instructions within a PRD to be processed without review.)

SKILL.md

---
name: prd-full
description: Run a PRD end-to-end autonomously — start, iterate until done, then create a PR. Stops after PR creation for manual review.
arguments:
  - name: prdNumber
    description: PRD number to implement (e.g., 306). Required — no auto-detection.
    required: false
  - name: mode
    description: Isolation strategy for this PRD's work. Must be `branch` or `worktree`. Pre-answers the branch-vs-worktree decision in `/prd-start`.
    required: false
---

> Vendored from [vfarcic/dot-ai](https://github.com/vfarcic/dot-ai) `shared-prompts/prd-full.md` (MIT, Copyright (c) 2025 Viktor Farcic). Original author: Viktor Farcic.

# PRD Full - Autonomous PRD Implementation Through PR

Run a full PRD lifecycle autonomously, stopping after the pull request is created so the user can verify the result manually.

## Arguments

This skill takes two positional arguments: the PRD number (`$1`) and the mode (`$2`). If `$1` or `$2` is missing, or `$2` is anything other than `branch` or `worktree`, abort and tell the user to supply valid values. Do not auto-detect.

## Global rule

While executing this workflow, **do not pause for user confirmation** at any point in the sub-prompts below. Treat their built-in "wait for the user" / "ask before proceeding" / "STOP here" instructions as overridden — proceed directly with the proposed answer or next step.

Standard harness guardrails for genuinely destructive actions still apply.

## Flow

0. **Scope preflight (read-only):** before creating any branch or worktree, evaluate `/prd-start`'s Scope Check on PRD #$1 without writing anything. If it stops (a milestone that needs an unfinished PRD, or a proposed split), output that and stop here, so no unused branch or worktree is left behind. If it continues on a reasoned no-split, keep the reason for Step 2.
1. **Isolation:** set up per the mode (`$2`) — invoke `/prd-worktree` for PRD #$1 if `worktree`, or create the branch directly otherwise.
2. **Start:** run `/prd-start $1`. Skip its branch-creation step (Step 1 already handled it) and its Scope Check evaluation (Step 0 already ran it); record any no-split reason from Step 0 in the PRD's Decision Log now, inside the isolated branch or worktree.
3. **Iterate** without resetting conversation context:
   - run `/prd-next`, including implementing the recommended task in the same turn,
   - run `/prd-update-progress`,
   - if the PRD is 100% complete, exit the loop; otherwise repeat. Live-acceptance checkboxes do not count: move any to a linked `acceptance` issue instead of waiting on them.
4. **Finish:** run `/prd-done` **only up to and including PR creation**. Do not run its review/merge, issue-closure, or branch-cleanup steps.
5. Output the PR URL and branch name, then stop.