SuperAPI AI Directory

paw-ps-research-to-brief

Office & productivity

View repository

Updated 10-16 03:16

About this skill

Research to Brief

Overview

A deterministic workflow that converts validated research outputs into a clean, structured product brief. This pipeline ensures no evidence or uncertainty is lost during synthesis.

Args: Accepts --headless / -H for non-interactive execution.

Output: Product brief document with preserved signals (evidence and uncertainties).


Pipeline

Stage 1: Gather Inputs

Load ./references/01-gather-inputs.md

Collect all relevant research files from the product workspace:

  • Discovery outputs (product-context.md)
  • Market intelligence (market-intelligence.md)
  • Audience intelligence (audience-intelligence.md)

Progression: Proceed when at least one input source is located.

Stage 2: Validate Completeness

Load ./references/02-validate-completeness.md

Assess whether inputs meet minimum thresholds for brief synthesis. Identify gaps and either proceed with available data or flag for additional research.

Progression: Proceed when minimum input requirements are met OR explicit decision to proceed with gaps documented.

Stage 3: Synthesize Brief

Load ./references/03-synthesize-brief.md

Create the structured product brief by consolidating validated inputs into a single document. Apply consistent formatting and ensure traceability to source materials.

Progression: Proceed when brief structure is complete with all required sections populated.

Stage 4: Preserve Signals

Load ./references/04-preserve-signals.md

Document key evidence sources and unresolved uncertainties. Create signal preservation notes that trace decisions back to research findings.

Progression: Complete when all signals are documented and appended to the brief or decision log.


On Activation

  1. Determine the product workspace root: {project-root}/.pawbytes/prodig-suites/
  2. Identify active product context from .pawbytes/prodig-suites/active-product.md if it exists
  3. If no active product is set and not in headless mode, prompt user to specify product slug
  4. Begin pipeline at Stage 1

Input Sources

Source File Path Required
Discovery .pawbytes/prodig-suites/products/{product-slug}/product-context.md Preferred
Market Research .pawbytes/prodig-suites/products/{product-slug}/market-intelligence.md Preferred
Audience Research .pawbytes/prodig-suites/products/{product-slug}/audience-intelligence.md Preferred

Minimum Requirement: At least one of the three input sources must be available.


Output Structure

.pawbytes/prodig-suites/products/{product-slug}/
├── product-brief.md              # Main brief document
├── signal-notes.md               # Evidence and uncertainty log
└── product-decisions.md          # Appended brief metadata (if exists)

Output Contract

Every execution delivers:

  • Action type: Brief synthesis from research inputs
  • Files saved:
    • product-brief.md — Structured product brief
    • signal-notes.md — Evidence and uncertainty preservation
  • Recommendations: Next steps (e.g., validate with stakeholders, initiate design phase)
  • File saved to: .pawbytes/prodig-suites/products/{product-slug}/

Headless Mode

When invoked with --headless or -H:

  1. Attempt to read active product from .pawbytes/prodig-suites/active-product.md
  2. If not found, scan for most recently modified product workspace
  3. Execute all stages without prompts
  4. Document any gaps or assumptions in signal notes
  5. Exit with status code 0 on success, 1 if minimum inputs not met

Quality Gates

Stage Gate Condition
Gather Inputs At least 1 input file located
Validate Completeness Minimum input threshold met OR documented decision to proceed
Synthesize Brief All required brief sections populated
Preserve Signals All evidence and uncertainties documented