How do I install the Comanda skill?

Run npx skillmds@latest add kbarbel640-del/comanda in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Comanda safe to use?

This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: CAUTION, Skill Scanner: PASS. Capability flags: docs only. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Comanda?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Comanda free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Comanda?

kbarbel640-del (@kbarbel640-del) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Comanda by kbarbel640-del — Agent Skill

Name: Comanda
Rating: 4.666666666666667 (3 reviews)
Author: kbarbel640-del

Comanda

Generate, visualize, and execute declarative AI pipelines using the comanda CLI. Use when creating LLM workflows from natural language, viewing workflow charts, editing YAML workflow files, or processing/running comanda workflows. Supports multi-model orchestration (OpenAI, Anthropic, Google, Ollama, Claude Code, Gemini CLI, Codex).

kbarbel640-del Updated Sep 8, 2026 1 repo stars

File contents

Comanda - Declarative AI Pipelines

Comanda defines LLM workflows in YAML and runs them from the command line. Workflows can chain multiple AI models, run steps in parallel, and pipe data through processing stages.

Installation

# macOS
brew install kris-hansen/comanda/comanda

# Or via Go
go install github.com/kris-hansen/comanda@latest

Then configure API keys:

comanda configure

Commands

Generate a Workflow

Create a workflow YAML from natural language:

comanda generate <output.yaml> "<prompt>"

# Examples
comanda generate summarize.yaml "Create a workflow that summarizes text input"
comanda generate review.yaml "Analyze code for bugs, then suggest fixes" -m claude-sonnet-4-20250514

Visualize a Workflow

Display ASCII chart of workflow structure:

comanda chart <workflow.yaml>
comanda chart workflow.yaml --verbose

Shows step relationships, models used, input/output chains, and validity.

Process/Execute a Workflow

Run a workflow file:

comanda process <workflow.yaml>

# With input
cat file.txt | comanda process analyze.yaml
echo "Design a REST API" | comanda process multi-agent.yaml

# Multiple workflows
comanda process step1.yaml step2.yaml step3.yaml

View/Edit Workflows

Workflow files are YAML. Read them directly to understand or modify:

cat workflow.yaml

Workflow YAML Format

Basic Step

step_name:
  input: STDIN | NA | filename | $VARIABLE
  model: gpt-4o | claude-sonnet-4-20250514 | gemini-pro | ollama/llama2 | claude-code | gemini-cli
  action: "Instruction for the model"
  output: STDOUT | filename | $VARIABLE

Parallel Execution

parallel-process:
  analysis-one:
    input: STDIN
    model: claude-sonnet-4-20250514
    action: "Analyze for security issues"
    output: $SECURITY

  analysis-two:
    input: STDIN
    model: gpt-4o
    action: "Analyze for performance"
    output: $PERF

Chained Steps

extract:
  input: document.pdf
  model: gpt-4o
  action: "Extract key points"
  output: $POINTS

summarize:
  input: $POINTS
  model: claude-sonnet-4-20250514
  action: "Create executive summary"
  output: STDOUT

Generate + Process (Meta-workflows)

create_workflow:
  input: NA
  generate:
    model: gpt-4o
    action: "Create a workflow that analyzes sentiment"
    output: generated.yaml

run_it:
  input: NA
  process:
    workflow_file: generated.yaml

Available Models

Run comanda configure to set up API keys. Common models:

Provider	Models
OpenAI	`gpt-4o`, `gpt-4o-mini`, `o1`, `o1-mini`
Anthropic	`claude-sonnet-4-20250514`, `claude-opus-4-20250514`
Google	`gemini-pro`, `gemini-flash`
Ollama	`ollama/llama2`, `ollama/mistral`, etc.
Agentic	`claude-code`, `gemini-cli`, `openai-codex`

Examples Location

See ~/clawd/comanda/examples/ for workflow samples:

agentic-loop/ - Autonomous agent patterns
claude-code/ - Claude Code integration
gemini-cli/ - Gemini CLI workflows
document-processing/ - PDF, text extraction
database-connections/ - DB query workflows

Troubleshooting

"model not configured": Run comanda configure to add API keys
Workflow validation errors: Use comanda chart workflow.yaml to visualize and check validity
Debug mode: Add --debug flag for verbose logging

---name: comandadescription: Generate, visualize, and execute declarative AI pipelines using the comanda CLI. Use when creating LLM workflows from natural language, viewing workflow charts, editing YAML workflow files, or processing/running comanda workflows. Supports multi-model orchestration (OpenAI, Anthropic, Google, Ollama, Claude Code, Gemini CLI, Codex).---# Comanda - Declarative AI PipelinesComanda defines LLM workflows in YAML and runs them from the command line. Workflows can chain multiple AI models, run steps in parallel, and pipe data through processing stages.## Installation```bash# macOSbrew install kris-hansen/comanda/comanda# Or via Gogo install github.com/kris-hansen/comanda@latest```Then configure API keys:```bashcomanda configure```## Commands### Generate a WorkflowCreate a workflow YAML from natural language:```bashcomanda generate <output.yaml> "<prompt>"# Examplescomanda generate summarize.yaml "Create a workflow that summarizes text input"comanda generate review.yaml "Analyze code for bugs, then suggest fixes" -m claude-sonnet-4-20250514```### Visualize a WorkflowDisplay ASCII chart of workflow structure:```bashcomanda chart <workflow.yaml>comanda chart workflow.yaml --verbose```Shows step relationships, models used, input/output chains, and validity.### Process/Execute a WorkflowRun a workflow file:```bashcomanda process <workflow.yaml># With inputcat file.txt | comanda process analyze.yamlecho "Design a REST API" | comanda process multi-agent.yaml# Multiple workflowscomanda process step1.yaml step2.yaml step3.yaml```### View/Edit WorkflowsWorkflow files are YAML. Read them directly to understand or modify:```bashcat workflow.yaml```## Workflow YAML Format### Basic Step```yamlstep_name: input: STDIN | NA | filename | $VARIABLE model: gpt-4o | claude-sonnet-4-20250514 | gemini-pro | ollama/llama2 | claude-code | gemini-cli action: "Instruction for the model" output: STDOUT | filename | $VARIABLE```### Parallel Execution```yamlparallel-process: analysis-one: input: STDIN model: claude-sonnet-4-20250514 action: "Analyze for security issues" output: $SECURITY analysis-two: input: STDIN model: gpt-4o action: "Analyze for performance" output: $PERF```### Chained Steps```yamlextract: input: document.pdf model: gpt-4o action: "Extract key points" output: $POINTSsummarize: input: $POINTS model: claude-sonnet-4-20250514 action: "Create executive summary" output: STDOUT```### Generate + Process (Meta-workflows)```yamlcreate_workflow: input: NA generate: model: gpt-4o action: "Create a workflow that analyzes sentiment" output: generated.yamlrun_it: input: NA process: workflow_file: generated.yaml```## Available ModelsRun `comanda configure` to set up API keys. Common models:| Provider | Models ||----------|--------|| OpenAI | `gpt-4o`, `gpt-4o-mini`, `o1`, `o1-mini` || Anthropic | `claude-sonnet-4-20250514`, `claude-opus-4-20250514` || Google | `gemini-pro`, `gemini-flash` || Ollama | `ollama/llama2`, `ollama/mistral`, etc. || Agentic | `claude-code`, `gemini-cli`, `openai-codex` |## Examples LocationSee `~/clawd/comanda/examples/` for workflow samples:- `agentic-loop/` - Autonomous agent patterns- `claude-code/` - Claude Code integration- `gemini-cli/` - Gemini CLI workflows- `document-processing/` - PDF, text extraction- `database-connections/` - DB query workflows## Troubleshooting- **"model not configured"**: Run `comanda configure` to add API keys- **Workflow validation errors**: Use `comanda chart workflow.yaml` to visualize and check validity- **Debug mode**: Add `--debug` flag for verbose logging

Frequently asked questions