orchestra-scaffold
Scaffold the .orchestra/ agent knowledge base in a project — creates folder structure, templates, README, and ADR-000.
What this skill does
# Scaffold .orchestra/
Create the `.orchestra/` agent knowledge base structure. This is mechanical work — no questions asked, just files. When done, run `/orchestra-roadmap` to define the project vision and milestones interactively.
## What It Creates
```
.orchestra/
├── README.md
├── adr/
│ └── ADR-000-the-score.md
├── work/
│ └── TEMPLATES/
│ ├── prd.md
│ └── spec.md
├── uml/
└── devlog/
└── {YYYY}-Q{N}/
```
## Steps
### 1. Check
- Determine target path from $ARGUMENTS (default: current working directory)
- If `.orchestra/` already exists → STOP and report what's there
- Note the project name: use the directory name (e.g. `basename` of the path)
### 2. Create or Update Project README
The project `README.md` (at the project root, not inside `.orchestra/`) is the human-facing entry point and the source of the orchestra Brief.
**If no `README.md` exists:**
Create one with a Brief section template:
```markdown
# {directory-name}
> {one-line description — fill this in}
## Brief
**Vision:** {what done looks like at the highest level — fill this in}
**Audience:** {who benefits — fill this in}
## Usage
{fill in after first milestone ships}
```
**If `README.md` exists and has no `## Brief` section:**
Append the Brief section to the end of the file:
```markdown
## Brief
**Vision:** {what done looks like at the highest level — fill this in}
**Audience:** {who benefits — fill this in}
```
**If `README.md` exists and already has a `## Brief` section:**
Leave it untouched. Report it as already present.
### 3. Create Structure
Create all directories: `adr/`, `work/TEMPLATES/`, `uml/`, `devlog/{YYYY}-Q{N}/`
### 4. Write .orchestra/README.md
Explain the folder to agents and humans:
```markdown
# .orchestra/
Agent knowledge base for this project. All agents read this folder before acting.
## Structure
- `roadmap.md` — The score. Project vision and milestone index.
- `adr/` — Architecture Decision Records. Decisions that affect future agents.
- `work/` — Per-work-item folders. Each contains a PRD, spec, and gherkin.
- `uml/` — Mermaid diagrams for architecture, workflows, and state machines.
- `devlog/` — Chronological journal of development sessions.
## The Loop
```
/orchestra-roadmap → define vision and milestones
/orchestra-plan → PRD → Spec → Gherkin for a work item
/orchestra-implement → execute an approved spec
/orchestra-review → validate implementation against spec
/orchestra-merge → merge and close the work item
```
## Rules
- Agents read the roadmap and relevant ADRs before acting on any work item
- Every work item has a PRD before a spec; every spec before implementation
- Nothing moves forward without explicit human approval at each gate
```
### 5. Write ADR-000
Write `.orchestra/adr/ADR-000-the-score.md`:
```markdown
---
id: ADR-000
status: accepted
created_on: {YYYY-MM-DD}
---
# ADR-000: The Score
## Decision
This project uses Orchestra — a software development lifecycle encoded for agents. PRDs are the unit of work. Every significant piece of work has a PRD before a spec, and a spec before implementation.
## Rationale
Without a written PRD, work drifts. Without a spec, agents have no contract to
execute against. Without Gherkin, there is no definition of done an agent can verify.
The `.orchestra/` folder is the shared knowledge base. Agents read it. Humans update
it at each gate. The score doesn't change mid-performance without a new ADR.
## Consequences
- No implementation without an approved spec
- No spec without an approved PRD
- All significant architectural decisions are recorded as ADRs
```
### 6. Write Templates
Write `.orchestra/work/TEMPLATES/prd.md`:
```markdown
---
ticket:
status: draft
created_on:
---
# {Title}
## Objective
{What does "done" look like? 1–2 sentences.}
## Success Criteria
- [ ] {Testable criterion}
## Context
{Why this matters. Which milestone it serves.}
## Materials
| Deliverable | Location | Status |
|-------------|----------|--------|
| | | Not Started |
## References
## Notes
```
Write `.orchestra/work/TEMPLATES/spec.md`:
```markdown
---
ticket:
status: draft
created_on:
---
# {Title}
> PRD: {path}
## Objective
{Restated from PRD}
## Approach
### Step 1: {name}
## Testing Strategy
### Unit Tests
### Integration Tests
### E2E Tests
## Deliverables
| File | Purpose | Status |
|------|---------|--------|
## Acceptance Criteria
### Functional
### Unit
### Integration
### E2E
## Dependencies
## Risks
| Risk | Mitigation |
|------|-----------|
```
### 6. Report
```
## .orchestra/ Scaffolded
### Created
- README.md [created with Brief section] ← or "Brief section added" / "Brief already present"
- .orchestra/README.md
- .orchestra/adr/ADR-000-the-score.md
- .orchestra/work/TEMPLATES/prd.md
- .orchestra/work/TEMPLATES/spec.md
- .orchestra/uml/
- .orchestra/devlog/{quarter}/
### Next Step
Fill in the Brief section in README.md, then run /orchestra-roadmap.
The Brief is the minimum needed to seed the roadmap without typing it twice:
## Brief
**Vision:** {what done looks like at the highest level}
**Audience:** {who benefits}
Vision and audience are stable from day one. Milestones are a planning output —
/orchestra-roadmap derives them from the vision in conversation with you.
/orchestra-roadmap will read the Brief and propose milestones for your approval.
If no Brief is present it will ask you the questions interactively.
```
Stop here. Do not ask about the project. Do not generate a roadmap.
Related in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts — single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score ≥ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.