Skip to content
Skills Library
AI & agents · Learning & explanation

Step Actions Reference

/Step Actions Reference

Reference for workflow step action types, parameters, and configuration. Use when building workflow.json steps, choosing between agent.run and agent.session, configuring git.clone options, repo.identify, ci.executeWorkflow, or setting workflow-level and step-level configuration like priority, timeoutMs, and stepMaxDurationMinutes.

Licence
Apache-2.0
Last repository change
16.09.2026

Assessment and security

90/ 100Class: good
Format15 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
2
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 16.09.2026, without references to retired models or APIs.

Licence

Apache-2.0, otwarta: można używać i zmieniać, z zachowaniem licencji i informacji o zmianach.

What it does and when to use it

Reference for workflow step action types, parameters, and configuration. Use when building workflow.json steps, choosing between agent.run and agent.session, configuring git.clone options, repo.identify, ci.executeWorkflow, or setting workflow-level and step-level configuration like priority, timeoutMs, and stepMaxDurationMinutes.

Original author description

Reference for workflow step action types, parameters, and configuration. Use when building workflow.json steps, choosing between agent.run and agent.session, configuring git.clone options, repo.identify, ci.executeWorkflow, or setting workflow-level and step-level configuration like priority, timeoutMs, and stepMaxDurationMinutes.

How it works

  1. Action Types Overview
  2. Step Structure
  3. When to Use Each Action
  4. Common Pattern: repo.identify → git.clone
  5. Workflow-Level Configuration
  6. Agent Engine
Original SKILL.md headings · 189 lines, 791 words
SKILL.md preview
---
name: Step Actions Reference
description: >
  Reference for workflow step action types, parameters, and configuration.
  Use when building workflow.json steps, choosing between agent.run and agent.session,
  configuring git.clone options, repo.identify, ci.executeWorkflow, or setting
  workflow-level and step-level configuration like priority, timeoutMs, and stepMaxDurationMinutes.
---

# Step Actions Reference

This skill covers how to configure steps in `workflow.json`, including all available action types, their parameters, and workflow/step-level configuration.

## Action Types Overview

| Action | Purpose | Has Instruction? | Key Params |
|--------|---------|-------------------|------------|
| `agent.run` | Single agent executes a prompt | Yes (required) | `agentId`, `agentEngine` |
| `agent.session` | Coordinator orchestrates agents | Yes (required) | `agentIds`, `goal`, `exitCriteria` |
| `git.clone` | Clone repositories | No | `repoFullName`, `branch`, `cloneOptions` |
| `repo.identify` | Identify relevant repos for a ticket | No | `maxResults`, `minConfidence` |
| `ci.executeWorkflow` | Trigger external CI pipeline | No | `repoFullName`, `workflowId`, `ref`, `inputs` |

Two internal actions (`agent.internal.cache-repo`, `agent.internal.index-repo`) exist but are not for playbook use.

## Step Structure

Every step follows this shape:

```json
{
  "id": "step-id",
  "name": "Human-Readable Name",
  "action": "agent.run",
  "params": { ... },
  "instruction": "Prompt content from .md file",
  "stepMaxDurationMinutes": 30
}
```

**Rules:**
- `id` must match `^[a-zA-Z0-9-]+$` (letters, numbers, hyphens only)
- `id` must match the corresponding `.md` prompt filename
- `instruction` is required for `agent.run` and `agent.session`, omitted for infrastructure steps
- `stepMaxDurationMinutes` (min: 1) limits individual step execution time

## When to Use Each Action

### `agent.run` — Single-pass tasks

Use for straightforward tasks where one agent can complete the work:
- Analyzing code and producing a plan
- Reading data and formatting output
- Simple file generation or modification

### `agent.session` — Coordinated execution

Use when a step needs oversight, multi-agent coordination, or iteration:
- **Multi-agent delegation**: Coordinator assigns chunks of work to sub-agents
- **Iterative loops**: Draft → review → revise cycles with max iteration limits
- **Error recovery**: Coordinator detects failures and retries or adjusts
- **Verification-heavy tasks**: Each sub-task's output must be confirmed before proceeding
- **Interactive sessions**: User provides feedback during execution (via `listenToComments`)

See the `agent-session-design` skill for detailed coordinator patterns.

### `git.clone` — Repository cloning

Clones one or more repositories into the agent workspace:

```json
{
  "id": "git-clone",
  "name": "Clone Repo",
  "action": "git.clone",
  "params": {
    "repoFullName": "{{trigger.repository.fullName}}",
    "branch": "{{trigger.pullRequest.headBranch}}",
    "cloneOptions": {
      "depth": 1,
      "filter": { "type": "blob:none" },
      "singleBranch": true
    }
  },
  "instruction": null
}
```

**`repoFullName`** accepts:
- A literal string: `"owner/repo"`
- A template expression: `"{{trigger.repository.fullName}}"` or `"{{outputs.identify-repos}}"`
- At runtime, can resolve to a string, string array, or object array (from `repo.identify`)

**`cloneOptions`**:
- `depth` (int): Shallow clone depth (1 = latest commit only)
- `filter.type` (string): Partial clone filter (`"blob:none"`, `"blob:limit=100M"`)
- `singleBranch` (bool): Clone only the specified branch
- `sparseCheckout` (object): `{ enabled: true, paths: ["src/", "docs/"] }`
- `ignoreCache` (bool): Force fresh clone
- `submodules` (object): `{ enabled: true, shallow: true }`

### `repo.identify` — Repository identification

Identifies the most relevant repositories for a ticket context (no code access):

```json
{
  "id": "identify-repos",
  "name": "Identify Repositories",
  "action": "repo.identify",
  "params": {
    "maxResults": 3,
    "minConfidence": 0.2
  },
  "instruction": null
}
```

- `maxResults` (1-20, default: 1): Maximum repos to return
- `minConfidence` (0-1, default: 0.2): Minimum confidence threshold
- `identificationHints` (string, optional): Free text to bias identification

**Output**: Referenced by downstream steps as `{{outputs.identify-repos}}`. Typically fed into `git.clone`'s `repoFullName`.

### `ci.executeWorkflow` — External CI pipeline

Triggers a CI/CD pipeline on an external repository:

```json
{
  "id": "run-tests",
  "name": "Run Test Suite",
  "action": "ci.executeWorkflow",
  "params": {
    "repoFullName": "{{trigger.repository.fullName}}",
    "workflowId": "test-suite.yml",
    "ref": "{{trigger.pullRequest.headBranch}}",
    "inputs": { "test_type": "integration" },
    "waitForCompletion": true
  }
}
```

- `repoFullName` (string): Repository identifier with provider inference
- `workflowId` (string): Workflow filename or pipeline ID
- `ref` (string, optional): Branch or tag reference
- `inputs` (Record<string, string>, optional): Key-value workflow inputs
- `waitForCompletion` (bool, optional): Whether to block until the pipeline finishes

## Common Pattern: repo.identify → git.clone

Many issue-triggered workflows need to identify which repos to work with:

```json
{
  "flow": [
    { "from": "", "to": "identify-repos" },
    { "from": "identify-repos", "to": "clone-repo" },
    { "from": "clone-repo", "to": "next-step" }
  ],
  "steps": [
    {
      "id": "identify-repos",
      "action": "repo.identify",
      "params": { "maxResults": 3, "minConfidence": 0.2 }
    },
    {
      "id": "clone-repo",
      "action": "git.clone",
      "params": {
        "repoFullName": "{{outputs.identify-repos}}",
        "cloneOptions": { "depth": 1 }
      }
    }
  ]
}
```

## Workflow-Level Configuration

These fields go in `workflow.definition`:

| Field | Type | Default | Description |
|-------|------|---------|-------------|
| `priority` | 1-100 | 5 | Lower = higher priority. Built-in playbooks use 1-10. |
| `timeoutMs` | number | — | Abort workflow after this time (min: 30000ms). |
| `statusUpdateMethod` | string | `"comment"` | How status updates appear: `"comment"` (new comment per execution), `"reuse_comment"` (reuse existing), `"static_comment"` (initial + completion only) |
| `defaultModelKey` | string | null | Default LLM model for all agent steps |

## Agent Engine

Both `agent.run` and `agent.session` support `agentEngine`:
- `"overcut"` (default): Full Overcut agent runtime
- `"claude"`: Direct Claude API execution

For the complete parameter schemas, see `references/action-params.md`.
Files in this package (2)
  • SKILL.md 7 KB
  • references/action-params.md 8 KB

Versions

The database has no other skills named Step Actions Reference.

Database snapshot: 13.11.2026 · package version d3b33c17f1a7 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 90agent-workflow
    Create, update, run, resume, or continue agent workflows. Use when the user asks to create or fix workflow JSON, execute an existing workflow, run a workflow with a prompt, resume a workflow run, or continue the next node in a waiting multi-step workflow.
  • 100execute
    Run the task's current workflow step: pick the agent, have a brief written, start the agent and record the result. Use when the person says "run it", "do the next step", "go ahead with the build", "carry on with T-02", or a plan is accepted and the step is a run, check or make-together step.
  • 85wp-clone-run
    The step detail of the /wp-clone run, split out of commands/wp-clone.md so the command stays a short map — Path A (SSH automated clone, A1-A13), Path B (manual import from a SQL dump and uploads archive, B1-B7), Path C (restore from a full-site files archive plus a database…
  • 100lindy-core-workflow-a
    Build and configure multi-step Lindy AI agent workflows. Use when creating agents with triggers, actions, conditions, knowledge bases, or agent steps. Trigger with phrases like "create lindy agent", "build lindy agent", "lindy agent workflow", "configure lindy agent", "lindy workflow".
  • 90github-actions
    Apply when writing or editing any workflow, composite action, job, or step — and read the owner pointers before adding a rule. Esposter's GitHub Actions authoring for `.github/workflows` and `.github/actions`: how a step is written.
  • 95design
    This skill should be used when the user has a Workflow Requirements document and wants to design an AI workflow. It gathers architecture decisions, assesses workflow autonomy level, chooses an orchestration mechanism and involvement mode, classifies steps, maps building blocks,…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.