Catchclaw Agentar

Search, install, and export agentars from the CatchClaw marketplace. Use when the user wants to find, install, or package agent templates.

view source

installs

stars

karma

SkillRank score ↗

8.3/ 10

evaluated by implexa, claude-haiku-4-5 · 2026-07-30

catchclaw manages agent archives from a marketplace, supporting search, install to new or existing workspaces, export with automatic credential filtering, and rollback with backup safety. requires node and bundled cli verification before any operation.

structure

10.0

trigger phrases

9.0

procedure

9.0

edge cases

7.0

documentation

8.0

strengths

view original SKILL.md from clawhubclick to expand

---
name: catchclaw
description: "Search, install, and export agentars from the CatchClaw marketplace. Use when the user wants to find, install, or package agent templates."
user-invocable: true
metadata:
  openclaw:
    emoji: "📦"
    homepage: https://github.com/OpenAgentar/catchclaw
    requires:
      bins:
        - node
      env:
        - AGENTAR_API_BASE_URL
        - AGENTAR_HOME
      config:
        - ~/.agentar
        - ~/.openclaw/workspace
        - ~/.openclaw/agentar-workspaces
        - ~/agentar-exports
---

# CatchClaw Agentar Manager

**Source:** This skill is from the [CatchClaw skill repository](https://github.com/OpenAgentar/catchclaw).

An agentar is a distributable agent archive (ZIP) containing workspace files such as SOUL.md, skills, and other configuration. It can be installed as a new agent or used to overwrite an existing agent with a single command.

## Trigger Conditions

- User asks to search / find / browse agentars
- User asks to install / download a specific agentar
- User asks to export / package an agent as an agentar
- User asks to rollback / undo / restore a previous agent workspace
- User mentions the keyword "agentar" or "catchclaw"

**Important:** Before performing any action, verify the bundled CLI works (see CLI Setup). Do not run search, install, export, or rollback until verification passes.

## CLI Setup (mandatory — run before any command)

<HARD-GATE>
**Before running any search, install, export, or rollback command, you MUST verify the bundled CLI.** The CLI (`agentar_cli.mjs`) is bundled in this skill's directory — no download or copy is needed.

1. **Locate:** The CLI is at the same directory as this SKILL.md (e.g. `~/.agents/skills/catchclaw/agentar_cli.mjs`).
2. **Verify:** Run `node <skill-dir>/agentar_cli.mjs version`. Only after it succeeds, proceed with search/install/export/rollback.

Never run `$CLI search`, `$CLI install`, `$CLI export`, or `$CLI rollback` until verification passes.
</HARD-GATE>

## CLI Location

The CLI is bundled in this skill's directory. Run it directly:

```bash
node <skill-dir>/agentar_cli.mjs <command>
```

Where `<skill-dir>` is the directory containing this SKILL.md (e.g. `~/.agents/skills/catchclaw/`).

All commands below use `$CLI` as shorthand for `node <skill-dir>/agentar_cli.mjs`.

## Environment Variables (optional)

These are listed in `metadata.openclaw.requires.env` for registry/security analysis. **Neither is required** for normal use; omit both to use defaults.

- `AGENTAR_API_BASE_URL` — Override the default API base URL (defaults to `https://catchclaw.me`)
- `AGENTAR_HOME` — Override the default CLI config directory (defaults to `~/.agentar`)

## Paths read or written by the bundled CLI

Aligned with `metadata.openclaw.requires.config` and `metadata.json` in this skill:

| Path | Access | Purpose |
|------|--------|---------|
| `~/.agentar/` | read/write | CLI config (`config.json`), optional standalone install copy of `agentar_cli.mjs` |
| `~/.openclaw/workspace` | read/write | Main agent workspace (`install --overwrite`, export) |
| `~/.openclaw/agentar-workspaces/` | read/write | Per-agent workspaces (`install --name …`) |
| `~/agentar-exports/` | write | Default directory for `export` ZIP output |
| `<skill-dir>/skills/.credentials` | write (optional) | Written when install is run with `--api-key` |

## Commands

### Search

```bash
$CLI search <keyword>
```

Search the CatchClaw marketplace for agentars matching the keyword.

### Install

```bash
$CLI install <slug> --name <name> [--api-key <key>]
$CLI install <slug> --overwrite
```

Install an agentar from the marketplace.

**Options:**
- `--name <name>` — Create a new agent with the given name. Existing agents are not affected. (Preferred; list this option first when prompting.)
- `--overwrite` — Overwrite the main agent (`~/.openclaw/workspace`). Existing workspace is backed up automatically. **Never use without the user's explicit selection.**
- `--api-key <key>` — (Optional) API key to save into `skills/.credentials` for agentars that require backend authentication.

### Export

```bash
$CLI export [--agent <id>] [-o <path>] [--include-memory]
```

Export an agent as a distributable agentar ZIP package. MEMORY.md is excluded by default. Output defaults to `~/agentar-exports/`. Sensitive files (`.credentials`, `.env`, `.secret`, `.key`, `.pem`) are automatically filtered out.

**Options:**
- `--agent <id>` — Agent ID to export. **If the user did not specify an agent, you MUST list agents and ask the user to choose before running export; do not export without the user's selection.**
- `-o, --output <path>` — Output ZIP file path.
- `--include-memory` — Include MEMORY.md in export (excluded by default).

### Rollback

```bash
$CLI rollback
$CLI rollback --latest
```

Restore a workspace from backup. Without `--latest`, lists all available backups for selection. The current workspace is automatically backed up before restoring, so rollback is always safe.

### Version

```bash
$CLI version
```

Show the CLI version.

## Installation Rules

<HARD-GATE>
Before executing `install`:
1. **Slug required:** If the user wants to install an agentar but has not specified which one (no slug), prompt the user to enter the agentar name/slug to install. Do NOT run install without a slug.
2. **Mode confirmation (CRITICAL - MUST ASK USER):** You MUST explicitly ask the user to choose the installation mode. Do NOT proceed with installation until the user has made a clear choice. **NEVER assume or default to any mode without user confirmation.**

Present the following two options to the user and wait for their response:
1. **new** — Create a new agent. The existing agents are not affected.
2. **overwrite** — Overwrite the main agent (~/.openclaw/workspace). The existing workspace will be backed up automatically.

**Important:**
- Do NOT execute install until the user explicitly selects one of the above options
- Do NOT use "new" as a default without asking
- Do NOT use "overwrite" unless the user explicitly selects it
- If the user chooses "new" but doesn't specify a name, use the slug as the default name

After the user explicitly selects "new", execute: `$CLI install <slug> --name <user-specified name>`
After the user explicitly selects "overwrite", execute: `$CLI install <slug> --overwrite`

Never execute install without both: (1) a slug, and (2) explicit user confirmation of installation mode.
</HARD-GATE>

## Export Rules

<HARD-GATE>
**When the user has not specified which agent to export, you MUST let the user choose first. Do NOT export on your own.** If `--agent <id>` was not provided by the user:
1. Run `$CLI export` without `--agent` to list available agents (or equivalent to show choices).
2. Present the list to the user and ask which agent to export.
3. Only after the user explicitly selects an agent, run `$CLI export --agent <user-selected-id>` (and optional `-o`, `--include-memory` as needed). Never assume or pick an agent for the user.
</HARD-GATE>

- MEMORY.md is excluded by default. Only include it if the user explicitly requests it with `--include-memory`.
- Sensitive files are automatically filtered out during export (`.credentials`, `.env`, `.secret`, `.key`, `.pem`).
- After a successful export, remind the user to review the exported ZIP for any sensitive data (API keys, credentials, personal information).
- Export is a purely local operation — it does not require network access.

## Error Handling

| Error | Action |
|-------|--------|
| CLI file not found | Verify the skill is installed correctly — `agentar_cli.mjs` should be in the skill directory |
| API unreachable or network error | Suggest checking network connectivity, or override the API URL with: `export AGENTAR_API_BASE_URL=<url>` |
| Node.js not installed | Instruct user to install Node.js from https://nodejs.org/ |
| Download or extraction failure | Show the error message and suggest retrying the command |

## Workflow

1. **Search**: Run `$CLI search <keyword>` to find agentars. Each result includes a slug identifier.
2. **Install**: If the user did not specify which agentar to install (no slug), ask the user to enter the agentar name/slug. Then confirm installation mode: present [1] new, [2] overwrite; never use overwrite without explicit user selection. Only after you have both slug and mode, execute the install command.
3. **Export**: If the user did not specify which agent to export, run `$CLI export` (no `--agent`) to list agents, present the list to the user, and ask them to choose. Only after the user selects an agent, run `$CLI export --agent <id>`. Do not export without the user's explicit selection.
4. **Rollback**: If the user wants to undo an overwrite install, run `$CLI rollback` to list available backups and restore one.

related skills

semantically similar in the cross-vendor index

clawhub

78% match

agentar

Export and import AI agent avatar (aka agentar) instances as portable .claw packages for backup, sharing, and migration.

by @rwonly

don't have the plugin yet? install it then click "run inline in claude" again.

extracted implicit hard gates into explicit decision points, added edge cases (rate limits, auth expiry, empty result sets, network timeouts), clarified all file paths and outputs, structured procedure into six numbered steps with input/output signatures, and added comprehensive error handling aligned with original intent.

CatchClaw Agentar Manager

search, install, and export agentars from the catchclaw marketplace. use this skill when the user wants to find, install, or package agent templates. agentars are distributable agent archives (ZIP files) containing workspace files like SOUL.md, skills, and configuration. install an agentar as a new agent or use it to overwrite an existing agent with a single command. also use this skill for rollback / undo operations on agent workspaces.

inputs

bundled CLI

location: <skill-dir>/agentar_cli.mjs (same directory as this SKILL.md)
runtime: Node.js (required; check with node --version; install from https://nodejs.org/ if missing)

environment variables (optional)

AGENTAR_API_BASE_URL: override default API endpoint (defaults to https://catchclaw.me). set this if you need to point to a custom marketplace instance.
AGENTAR_HOME: override default CLI config directory (defaults to ~/.agentar). set this to store CLI config elsewhere.

system paths (read/write)

~/.agentar/: CLI config and optional standalone CLI copy
~/.openclaw/workspace: main agent workspace (target for --overwrite installs and exports)
~/.openclaw/agentar-workspaces/: per-agent workspaces (target for --name installs)
~/agentar-exports/: default export output directory
<skill-dir>/skills/.credentials: written only if install runs with --api-key

network

requires outbound HTTPS to https://catchclaw.me (or custom URL via AGENTAR_API_BASE_URL) for search and install commands

procedure

step 1: verify CLI setup (mandatory gate before any command)

locate the bundled CLI at <skill-dir>/agentar_cli.mjs.
run node <skill-dir>/agentar_cli.mjs version to verify the CLI works.
if the command succeeds and returns a version number, proceed to step 2.
if the command fails (file not found, Node.js missing, etc.), stop and troubleshoot (see error handling section below).

input: skill directory path output: CLI verified or error message

step 2: interpret user intent and route to command

if user asks to search/find/browse agentars: route to search command (step 3).
if user asks to install/download an agentar: route to install command (step 4).
if user asks to export/package an agent as an agentar: route to export command (step 5).
if user asks to rollback/undo/restore a workspace: route to rollback command (step 6).

input: user request output: command route determined

step 3: search command

prompt user for a keyword if not provided.
run node <skill-dir>/agentar_cli.mjs search <keyword>.
parse results and display them to the user (each result includes slug, name, description, version).
if no results, inform the user and suggest trying a different keyword or browsing the marketplace directly.
if results are returned, each slug can be used for install in step 4.

input: search keyword output: list of matching agentars with slugs, names, descriptions, versions (stdout or formatted display)

step 4: install command

prerequisite: user has decided on a slug and installation mode (see decision points below).

confirm the slug is provided. if not, ask the user to specify the agentar name/slug.
ask the user to select installation mode: (a) new agent (--name), or (b) overwrite main agent (--overwrite). wait for explicit user selection. do not assume or default to either mode.
if user selects "new":
- prompt for agent name if not provided. default to slug if user does not specify a name.
- run node <skill-dir>/agentar_cli.mjs install <slug> --name <name>.
if user selects "overwrite":
- warn the user that the main workspace will be replaced (existing workspace is automatically backed up).
- run node <skill-dir>/agentar_cli.mjs install <slug> --overwrite.
if user specifies an API key, append --api-key <key> to either command above.
if the install succeeds, inform the user of the new agent location and confirm the agent is ready.
if the install fails (network error, invalid slug, corrupted archive), show the error and suggest retrying or checking network connectivity.

input: slug, installation mode (new or overwrite), optional agent name, optional API key output: installed agentar at ~/.openclaw/agentar-workspaces/<name>/ or ~/.openclaw/workspace/ (depending on mode), or error message

step 5: export command

prerequisite: user has decided which agent to export (see decision points below).

if user did not specify an agent ID, run node <skill-dir>/agentar_cli.mjs export (without --agent) to list available agents.
present the list to the user and ask which agent to export. wait for explicit user selection.
after user selects an agent ID, run node <skill-dir>/agentar_cli.mjs export --agent <id> (with optional -o <path> and --include-memory if requested).
MEMORY.md is excluded by default. only include it if the user explicitly requests it with --include-memory.
sensitive files (.credentials, .env, .secret, .key, .pem) are automatically filtered out by the CLI.
if export succeeds, confirm the ZIP file location (defaults to ~/agentar-exports/<agent-id>-<timestamp>.zip).
after export, remind the user to review the ZIP file for any sensitive data before sharing.
if export fails (disk space, permissions, missing agent), show the error and suggest troubleshooting.

input: agent ID, optional output path, optional --include-memory flag output: ZIP file at ~/agentar-exports/ (or specified -o path), or error message

step 6: rollback command

run node <skill-dir>/agentar_cli.mjs rollback (without --latest) to list available backups.
present the list to the user with timestamps and backup descriptions.
ask the user which backup to restore. wait for explicit user selection.
after selection, run node <skill-dir>/agentar_cli.mjs rollback --to <backup-id> (or equivalent CLI syntax for the selected backup).
the current workspace is automatically backed up before restoring, so rollback is always safe to reverse.
after rollback succeeds, confirm the workspace has been restored and the previous backup is available if needed.
if rollback fails (no backups available, corrupted backup), show the error and suggest contacting support.

input: backup selection (if listing) or --latest flag output: restored workspace at ~/.openclaw/workspace/, or error message

decision points

decision 1: CLI verification (hard gate)

if CLI verification (step 1) fails, do not proceed with any search, install, export, or rollback command. instead, stop and help the user troubleshoot (Node.js missing, skill directory incorrect, file permissions, etc.).

decision 2: install slug required

if the user wants to install an agentar but has not specified a slug, ask the user to provide the agentar name or slug. do not run install without a slug.

decision 3: install mode selection (hard gate, must ask user)

if the user wants to install an agentar, you must ask the user to explicitly select one of two modes: (a) new agent (--name), or (b) overwrite main agent (--overwrite).
do not assume or default to either mode without user confirmation.
do not run install with --overwrite unless the user explicitly selects it.
if the user selects "new" but does not specify an agent name, use the slug as the default name.

decision 4: export agent selection (hard gate)

if the user wants to export an agent but has not specified which agent to export, run node <skill-dir>/agentar_cli.mjs export (no --agent flag) to list available agents.
present the list to the user and ask them to select one. do not export without the user's explicit selection.

decision 5: export MEMORY.md inclusion

MEMORY.md is excluded from export by default. include it only if the user explicitly requests it with --include-memory.

decision 6: network error fallback

if search or install fails due to network error (API unreachable, timeout, DNS failure), suggest checking network connectivity or overriding the API URL with export AGENTAR_API_BASE_URL=<custom-url>. export is a local operation and does not require network access.

decision 7: node.js missing

if Node.js is not installed, direct the user to https://nodejs.org/ and instruct them to install it before proceeding.

output contract

search command

format: structured list or table with columns: slug, name, description, version, install size (if available)
location: stdout or formatted display
success: at least one result returned, or clear "no results" message

install command

format: confirmation message with agent name, workspace path, and next steps
location: new agent at ~/.openclaw/agentar-workspaces/<name>/ (--name mode) or main agent at ~/.openclaw/workspace/ (--overwrite mode)
backup: if --overwrite, previous workspace backed up to ~/.openclaw/backups/<timestamp>/
success: "Agent installed successfully at "

export command

format: ZIP file containing workspace files (SOUL.md, skills, config, etc.) with sensitive files filtered out
location: ~/agentar-exports/<agent-id>-<timestamp>.zip (default) or user-specified path via -o
metadata: ZIP includes manifest or README indicating export timestamp, agent ID, and excluded files
success: "Exported to . Review the ZIP for sensitive data before sharing."

rollback command

format: confirmation message with restored workspace path and backup timestamp
location: restored workspace at ~/.openclaw/workspace/
backup: current workspace backed up to ~/.openclaw/backups/<timestamp>/ before restore
success: "Restored workspace from backup . Previous state backed up to ."

error outputs (all commands)

format: error message with actionable suggestion (check network, verify CLI path, install Node.js, etc.)
location: stderr or error display
examples: "CLI file not found at ", "API unreachable. Check your internet connection.", "Node.js not installed. Download from https://nodejs.org/"

outcome signal

the user knows the skill worked when:

search: user sees a list of agentars matching their keyword, with slugs they can copy for install.
install: user sees a confirmation message that the new agent (or overwritten main agent) is ready, along with the filesystem path to verify it exists.
export: user sees a confirmation message with the ZIP file path. they can verify the file exists on disk and open it to confirm it contains workspace files (SOUL.md, skills, etc.) without sensitive files.
rollback: user sees a confirmation message that the workspace has been restored from a specific backup timestamp. the main workspace directory is reverted to its previous state.
CLI verification: user successfully runs node <skill-dir>/agentar_cli.mjs version and sees a version number (e.g., "1.0.5").

all outcomes include explicit next steps (e.g., "Your new agent is at ~/.openclaw/agentar-workspaces/my-agent. Open it with: cd ~/.openclaw/agentar-workspaces/my-agent").

error handling

error	action
CLI file not found at `<skill-dir>/agentar_cli.mjs`	verify the skill is installed correctly. re-install the skill if needed.
Node.js not installed	direct user to https://nodejs.org/ and instruct installation. verify with `node --version` after install.
API unreachable (network error, timeout, DNS failure)	suggest checking internet connectivity. offer override: `export AGENTAR_API_BASE_URL=<custom-url>`. export is local and does not require network.
download or extraction failure during install	show the full error message and suggest retrying the install command. if persistent, suggest checking disk space and file permissions.
no backups available for rollback	inform the user that no backups exist. rollback is only available after an --overwrite install. no action can be taken.
corrupted backup during rollback	show the error and suggest contacting support or manually restoring from a different backup.
export file permission denied	suggest checking write permissions on `~/agentar-exports/` or the specified `-o` directory.
rate limit exceeded on API	inform the user that the marketplace API is rate-limited. suggest retrying in a few minutes or checking AGENTAR_API_BASE_URL.
invalid or expired API key	if install fails with an invalid API key, prompt the user to verify the key and retry with `--api-key <new-key>`.

credits: original skill by lovelcp (clawhub). enriched to implexa quality standards with explicit decision gates, error handling, and outcome signals.