Commands System
Relevant source files
The following files were used as context for generating this wiki page:
The Commands System in Everything Claude Code (ECC) provides a structured interface for user-invoked actions through a "slash-command" syntax. While the project is transitioning toward a skills-first architecture, the commands system remains a critical compatibility and workflow layer, hosting 92 specialized slash commands that orchestrate complex multi-step processes like code reviews, documentation updates, and implementation planning.
Sources: AGENTS.md:3-5, AGENTS.md:162, CONTRIBUTING.md:214-216
Architecture and Structure
Commands are defined as Markdown files located within the commands/ directory. Each file serves as a specification for a specific workflow, detailing its purpose, usage, and logic.
Command Component Layout
A standard command definition consists of three primary parts:
- Frontmatter: Metadata such as the description shown in help menus.
- Workflow Definition: Step-by-step logic the system follows when the command is triggered.
- Argument Handling: Instructions on how to process input variables (e.g.,
$ARGUMENTS).
Sources: CONTRIBUTING.md:218-243, scripts/ci/validate-commands.js:18-53
The following diagram illustrates the relationship between the command definition files and the broader ECC ecosystem:
The diagram shows how a slash command acts as an entry point to orchestrate agents and tools.
Sources: AGENTS.md:148-160, commands/ecc-guide.md:27-38
Core Command Categories
Commands are categorized based on their functional domain, often delegating heavy lifting to specialized agents.
| Category | Primary Commands | Associated Agents/Tools |
|---|---|---|
| Development | /code-review, /tdd, /build-fix | code-reviewer, tdd-guide, build-error-resolver |
| Planning | /plan, /multi-plan | planner, architect, ccg-workflow |
| Documentation | /update-docs, /ecc-guide | doc-updater, jsdoc2md |
| Automation | /commit, /e2e, /security-scan | git, playwright, security-reviewer |
Sources: AGENTS.md:34-70, CONTRIBUTING.md:245-253, commands/update-docs.md:13-20
Multi-Model Command Workflows
Advanced commands like /multi-plan or /backend utilize a multi-model approach, coordinating between different LLM backends (Codex, Gemini, and Claude) to perform specialized tasks.
Planning Workflow Example
The /multi-plan command follows a strict phased execution to ensure technical accuracy and avoid unauthorized code modifications.
This sequence highlights the parallel processing and orchestration role of Claude within the commands system.
Sources: commands/multi-plan.md:10-25, commands/multi-plan.md:130-165
Validation and Maintenance
To ensure system integrity, ECC includes a validation script (scripts/ci/validate-commands.js) that performs static analysis on the command files.
Validation Checks:
- Frontmatter Integrity: Ensures blocks are correctly delimited and formatted as valid YAML sequences or mappings.
- Cross-References: Verifies that references to other commands (e.g.,
`/build-fix`) or agents (e.g.,agents/planner.md) point to files that actually exist. - Workflow Integrity: Scans workflow diagrams within the Markdown to ensure referenced agents are valid.
- Content Sanity: Checks that command files are readable and non-empty.
Sources: scripts/ci/validate-commands.js:1-15, scripts/ci/validate-commands.js:95-144
Implementation Example: /update-docs
The /update-docs command illustrates how a command interacts with the filesystem to maintain project state. It identifies "Sources of Truth" and regenerates documentation automatically.
### Step 1: Identify Sources of Truth
| Source | Generates |
|--------|-----------|
| package.json | Available commands reference |
| .env.example | Environment variable documentation |
| Dockerfile | Infrastructure setup docs |
Sources: commands/update-docs.md:13-20
Legacy Compatibility and Future Direction
While commands/ remains a primary surface for slash-command entry, the project documentation specifies that skills/ is now the canonical workflow surface. Commands are increasingly viewed as "compatibility shims" to maintain cross-harness parity or provide legacy support for users who prefer slash-command interaction.
Sources: AGENTS.md:162-164, skills/ecc-guide/references/summary.md:104-110
Conclusion
The Commands System serves as the primary orchestration layer for Everything Claude Code. By combining simple Markdown specifications with advanced multi-model workflows and rigorous CI validation, it provides a reliable and extensible way for users to execute complex software engineering tasks. Although the project is moving towards a skills-based architecture, the Commands System continues to provide the essential structure required for high-level agent coordination and automated development processes.
Sources: AGENTS.md:3-8, CONTRIBUTING.md:214-216