πΌ Feature: Pre-built Artifact Templates (5β10 Ready-to-Use Patterns)
Context
The current pattern system (22 patterns in patterns/) is comprehensive but requires the user to compose their own prompts. They must understand what the pattern does, what inputs it needs, and how to trigger it. This creates friction: the user needs to read the pattern documentation before they can use it effectively.
Pre-built templates solve this by bundling the pattern with its data structure: the user fills in a simple data object (YAML) and gets a complete, quality-verified artifact without writing a single prompt.
Problem
- Using a pattern requires understanding the pattern's purpose and composition approach
- The user must write a custom prompt for every artifact, even for very common use cases
- Common workflows (e.g., "decision deck for two options", "project recap at end of sprint") are not single-command
- The gap between "I know what I want" and "I generated a complete artifact" is too large for casual use
What This Resolves
10 pre-built templates that can be invoked with a single command and a data file:
decision-deck β Compare 2β4 options with scores, risks, recommendations
project-recap β Sprint or milestone recap with progress, blockers, next steps
diff-review β PR/change review with annotated files, inline comments, approval signals
architecture-review β System/module architecture with connections, boundaries, trade-offs
audit-report β Security or process audit with findings, severity, actions
incident-postmortem β Incident timeline, root cause, learnings, action items
workflow-flowchart β Process workflow with steps, decision nodes, and roles
module-map β Codebase module map with dependencies, ownership, data flow
research-comparison β Options comparison with evidence, scores, trade-offs
ticket-triage β Issue triage board with priority, category, owner assignments
Each template has a single-file YAML input (human-editable, no code required), generates a complete quality-verified HTML artifact via deliver-artifact.py, and includes validation of the YAML input before generation.
Proposed Approach
Template structure (templates/prebuilt/<name>.yaml):
---
name: decision-deck
version: "1.0"
description: "Compare 2β4 options and recommend the best next move"
input_schema:
title:
type: string
required: true
options:
type: array
required: true
min_items: 2
max_items: 4
items:
name: string
score: integer # 0-100
risks: list[string]
evidence: list[string]
recommendation:
type: string
required: true
next_action:
type: string
required: true
Template engine (scripts/generate-template.py):
python scripts/generate-template.py decision-deck /path/to/data.yaml
# Generates: ~/.claude/html-explainer/outputs/artifacts/temporal/decision-deck-<hash>.html
# Runs deliver-artifact.py at 90+
python scripts/generate-template.py validate decision-deck /path/to/data.yaml
# Validates YAML against schema, returns errors if schema violated
CI / smoke tests: Each template has a canonical example in templates/prebuilt/examples/<name>-example.yaml. validate-templates.sh (new) runs all examples through the generator and verifies all pass deliver-artifact.py at 90+.
Acceptance Criteria
- At least 10 templates in
templates/prebuilt/ directory, each with YAML schema and example
scripts/generate-template.py accepts template name and data YAML, generates valid artifact
generate-template.py validate <template> <data.yaml> returns schema errors before generation
- All 10 templates generate artifacts that pass
deliver-artifact.py at 90+ using their example data
validate-templates.sh (new) runs all template examples and asserts all pass 90+
/make-the-right-html command recognizes template invocation patterns in natural language
- Template inputs are validated before generation (invalid YAML β clear error, no partial output)
- Documentation: each template file has a one-paragraph description and usage example
- New templates can be added to
templates/prebuilt/ without modifying the generator script
Quality Verification
Run validate-templates.sh to confirm all 10 templates generate artifacts at 90+. Verify generate-template.py list-templates shows all templates without hardcoding.
Labels
enhancement, templates, usability, pre-built
Milestone
v1.1 β Collaboration-ready artifacts
πΌ Feature: Pre-built Artifact Templates (5β10 Ready-to-Use Patterns)
Context
The current pattern system (22 patterns in
patterns/) is comprehensive but requires the user to compose their own prompts. They must understand what the pattern does, what inputs it needs, and how to trigger it. This creates friction: the user needs to read the pattern documentation before they can use it effectively.Pre-built templates solve this by bundling the pattern with its data structure: the user fills in a simple data object (YAML) and gets a complete, quality-verified artifact without writing a single prompt.
Problem
What This Resolves
10 pre-built templates that can be invoked with a single command and a data file:
decision-deckβ Compare 2β4 options with scores, risks, recommendationsproject-recapβ Sprint or milestone recap with progress, blockers, next stepsdiff-reviewβ PR/change review with annotated files, inline comments, approval signalsarchitecture-reviewβ System/module architecture with connections, boundaries, trade-offsaudit-reportβ Security or process audit with findings, severity, actionsincident-postmortemβ Incident timeline, root cause, learnings, action itemsworkflow-flowchartβ Process workflow with steps, decision nodes, and rolesmodule-mapβ Codebase module map with dependencies, ownership, data flowresearch-comparisonβ Options comparison with evidence, scores, trade-offsticket-triageβ Issue triage board with priority, category, owner assignmentsEach template has a single-file YAML input (human-editable, no code required), generates a complete quality-verified HTML artifact via
deliver-artifact.py, and includes validation of the YAML input before generation.Proposed Approach
Template structure (
templates/prebuilt/<name>.yaml):Template engine (
scripts/generate-template.py):CI / smoke tests: Each template has a canonical example in
templates/prebuilt/examples/<name>-example.yaml.validate-templates.sh(new) runs all examples through the generator and verifies all passdeliver-artifact.pyat 90+.Acceptance Criteria
templates/prebuilt/directory, each with YAML schema and examplescripts/generate-template.pyaccepts template name and data YAML, generates valid artifactgenerate-template.py validate <template> <data.yaml>returns schema errors before generationdeliver-artifact.pyat 90+ using their example datavalidate-templates.sh(new) runs all template examples and asserts all pass 90+/make-the-right-htmlcommand recognizes template invocation patterns in natural languagetemplates/prebuilt/without modifying the generator scriptQuality Verification
Run
validate-templates.shto confirm all 10 templates generate artifacts at 90+. Verifygenerate-template.py list-templatesshows all templates without hardcoding.Labels
enhancement, templates, usability, pre-built
Milestone
v1.1 β Collaboration-ready artifacts