agentic/code/addons/aiwg-utils/skills/scaffold-framework/SKILL.md
Create a complete framework package structure inside agentic/code/frameworks/
npx skillsauth add jmagly/aiwg scaffold-frameworkInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Create a complete framework package structure inside agentic/code/frameworks/.
Alternate expressions and non-obvious activations (primary phrases are matched automatically from the skill description):
| Pattern | Example | Action |
|---------|---------|--------|
| Named scaffold | "scaffold framework legal-ops" | Scaffold directly |
| Interactive | "scaffold framework --interactive" | Full guided design mode |
| Described | "create a framework for content production pipelines" | Derive name=content-production, confirm |
| Phased | "scaffold framework data-science --phases discovery,analysis,modeling,deployment" | Scaffold with custom phases |
Frameworks are major capability bundles that model complete operational lifecycles. They contain everything a team needs to manage a specific type of work end-to-end. Examples:
| Framework | Purpose | Scale |
|-----------|---------|-------|
| sdlc-complete | Software development lifecycle | ~58 agents, 100+ templates |
| media-marketing-kit | Marketing operations | ~37 agents, campaign lifecycle |
| forensics-complete | Digital forensics and IR | Investigation lifecycle |
| research-complete | Research workflow automation | Research phases |
Frameworks differ from addons:
Framework creation is a significant undertaking. Study existing frameworks before building new ones. Start with core agents per phase, then expand.
Extract from $ARGUMENTS:
<name> — kebab-case framework name (required)--description "<text>" — short description (optional; prompted if absent)--phases "<p1,p2,...>" — comma-separated phase list (optional; defaults to inception,elaboration,construction,transition)--author "<name>" — author name (optional)--interactive — enable guided design questionsIf <name> is missing, ask before proceeding.
ls agentic/code/frameworks/
This is the recommended path for first-time framework creation. Ask:
aiwg scaffold-framework <name> \
--description "<description>" \
--phases "<phase1,phase2,...>"
manifest.json — Framework registry entry:
{
"name": "<name>",
"version": "1.0.0",
"description": "<description>",
"author": "<author>",
"phases": ["<phase1>", "<phase2>"],
"agents": [],
"commands": [],
"skills": [],
"rules": [],
"templates": {}
}
README.md — Framework overview: purpose, phases, agent catalog, quick start.
actors-and-templates.md — Maps roles to artifacts per phase. Critical for orchestration.
plan-act-<name>.md — Master plan-and-act document defining framework behavior.
Each phase gets a flow definition in flows/:
# <Phase> Phase Flow
## Objectives
[What this phase accomplishes]
## Entry Criteria
[What must be true to begin this phase]
## Activities
[What happens in this phase]
## Artifacts Produced
[Documents and deliverables]
## Exit Criteria / Gate
[What must be true to leave this phase]
## Agents
[Which agents are active in this phase]
After scaffold, build agents in this order:
aiwg add-agent <role> --to <name> --template complex
flow-<phase> commandsflow-<phase>-to-<next> commandsproject-status, gate-check, health-checkaiwg add-command flow-<phase> --to <name> --template orchestration
# Deploy framework to active provider
aiwg use <name>
# Validate structure and metadata
aiwg validate-metadata
# Health check
aiwg doctor
agentic/code/frameworks/<name>/
├── README.md # Framework documentation
├── manifest.json # Framework configuration and registry
├── plan-act-<name>.md # Master plan-and-act document
├── actors-and-templates.md # Role-to-artifact mapping
├── agents/ # Agent definitions (.md files)
├── skills/ # Skill definitions
├── commands/ # Slash command definitions
├── rules/ # Framework-specific rules
├── templates/ # Document templates (by category)
├── flows/ # Phase flow definitions
│ ├── <phase1>.md
│ ├── <phase2>.md
│ └── ...
└── docs/ # Framework documentation
├── orchestrator-architecture.md
└── agent-catalog.md
Framework Created: <name>
─────────────────────────
Location: agentic/code/frameworks/<name>/
Phases: <phase1> → <phase2> → ... → <phaseN>
Created:
✓ README.md
✓ manifest.json
✓ plan-act-<name>.md
✓ actors-and-templates.md
✓ agents/
✓ skills/
✓ commands/
✓ rules/
✓ templates/
✓ flows/<phase>.md (for each phase)
✓ docs/
Next Steps:
1. Define actors: Edit actors-and-templates.md
2. Add agents: aiwg add-agent <role> --to <name> (2-3 per phase)
3. Add commands: aiwg add-command flow-<phase> --to <name>
4. Add templates: aiwg add-template <artifact> --to <name> --category <phase>
5. Deploy: aiwg use <name>
6. Validate: aiwg validate-metadata
Reference frameworks:
- sdlc-complete: 58 agents, ~100 templates (comprehensive)
- media-marketing-kit: 37 agents, marketing-focused
User: "scaffold framework legal-ops"
Action:
aiwg scaffold-framework legal-ops \
--description "Legal operations and case management lifecycle" \
--phases "intake,investigation,analysis,drafting,review,close"
Result: agentic/code/frameworks/legal-ops/ created with six phase flow files and full directory structure.
User: "scaffold framework --interactive"
Process: Guided questions establish the lifecycle type (e.g., content production), phases (ideation, research, writing, editing, publication), agent categories, and initial agent/command stubs.
User: "create a new AIWG framework called data-science-ops"
Action:
aiwg scaffold-framework data-science-ops \
--description "Data science project lifecycle management" \
--phases "discovery,data-preparation,modeling,evaluation,deployment"
data-ai
Report which research-corpus radar sidecars are overdue for refresh. Computes staleness (days since last refresh vs the cadence window) for every radar, sorted most-overdue-first. Runs via `aiwg corpus radar-status`.
data-ai
Aggregate research-corpus radar sidecars into a corpus or per-cluster freshness report — totals, overdue count, per-cluster / per-GRADE / per-trajectory breakdowns, an overdue table, and per-radar rationale snippets. Runs via `aiwg corpus radar-report`.
testing
Scaffold radar/freshness sidecars for research-corpus REFs. Pulls title/authors from the citation sidecar and GRADE from the analysis doc, defaults the refresh cadence from GRADE and the cluster from a corpus-local map, and stamps documentation/radar/REF-XXX-radar.md. Runs via `aiwg corpus radar-init`.
data-ai
Compute an entity's publication trajectory — per-year paper counts, topic drift, hot-streak detection (≥3 consecutive A-grade years), and career phase. Runs via `aiwg corpus profile-temporal`.