How do I install the Agents skill?

Run npx skillmds@latest add abzhaw/agents 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 Agents safe to use?

This skill has not completed SkillMD's automated safety review yet. Capability flags: makes network calls. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Agents?

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 Agents free to use?

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

Who published Agents?

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

Agents by abzhaw — Agent Skill

Name: Agents
Rating: 4.2 (5 reviews)
Author: abzhaw

File contents

Agents Skill – Creating & Updating Agent Manifests

Use this skill whenever you need to onboard a new agent or modify an existing one under agents/.

Checklist Before Editing

Read AGENTS.md (session rules) and the current memory/YYYY-MM-DD.md.
Identify which skills the agent needs (each must exist under skills/<name>/).
Confirm transport requirements (MCP endpoint, direct channel, etc.).

Creating a New Agent

Copy a template:

cp agents/juliamcp.yml agents/<new-agent>.yml

Set metadata:
- name: short handle (used in UI)
- kind: usually mcp-agent
- metadata.description: one-line summary

Configure transport:

transport:
  type: mcp
  endpoint: http://127.0.0.1:3001

Adjust endpoint/transport as needed.

Wire skills:

skills:
  required:
    - julia-bridge/send
    - julia-bridge/receive
    - julia-bridge/health

Behaviors: define retries, correlation IDs, pre-flight health checks, logging.
Logging: point to a file in logs/ if you need persistent traces.

Updating an Existing Agent

Locate the YAML under agents/.
Make the edit (add/remove skills, adjust behaviors, change endpoint).
Re-run any dependent MCP server to pick up config changes.
Test via the OpenClaw UI or CLI (trigger the agent and watch logs).
Record the change in memory/YYYY-MM-DD.md.

Validation Steps

yamllint agents/<agent>.yml (if available).
Trigger the agent once; ensure required skills resolve (no “skill not found” errors).
If agent uses MCP, run mcporter call <server>.bridge_health first.

Common Patterns

Health preflight: run a health skill before sending user messages.
Retry: allow one automatic retry with a short delay for MCP send failures.
Correlation IDs: enforce a prefix per agent (correlationIdPrefix).

Troubleshooting

Issue	Fix
Agent missing in UI	File not saved in `agents/` or YAML invalid – re-open and validate.
Skill not found	Ensure the skill folder + tool YAML exist, and the name matches.
MCP errors	Bridge down or config mismatch – follow `skills/mcp/SKILL.md`.
Duplicate correlation IDs	Adjust `correlationIdPrefix` or generate UUIDs.

Related Skills

skills/agents-system – bigger-picture maintenance
skills/mcp – MCP server operations
skills/julia-bridge – specific to the Julia bridge

---name: agentsdescription: Design, create, and update OpenClaw agent manifests in `agents/` (skills, transports, behaviors).---# Agents Skill – Creating & Updating Agent ManifestsUse this skill whenever you need to onboard a new agent or modify an existing one under `agents/`.## Checklist Before Editing- Read `AGENTS.md` (session rules) and the current `memory/YYYY-MM-DD.md`.- Identify which skills the agent needs (each must exist under `skills/<name>/`).- Confirm transport requirements (MCP endpoint, direct channel, etc.).## Creating a New Agent1. **Copy a template:** ```bash cp agents/juliamcp.yml agents/<new-agent>.yml ```2. **Set metadata:** - `name`: short handle (used in UI) - `kind`: usually `mcp-agent` - `metadata.description`: one-line summary3. **Configure transport:** ```yaml transport: type: mcp endpoint: http://127.0.0.1:3001 ``` Adjust endpoint/transport as needed.4. **Wire skills:** ```yaml skills: required: - julia-bridge/send - julia-bridge/receive - julia-bridge/health ```5. **Behaviors:** define retries, correlation IDs, pre-flight health checks, logging.6. **Logging:** point to a file in `logs/` if you need persistent traces.## Updating an Existing Agent1. Locate the YAML under `agents/`.2. Make the edit (add/remove skills, adjust behaviors, change endpoint).3. Re-run any dependent MCP server to pick up config changes.4. Test via the OpenClaw UI or CLI (trigger the agent and watch logs).5. Record the change in `memory/YYYY-MM-DD.md`.## Validation Steps- `yamllint agents/<agent>.yml` (if available).- Trigger the agent once; ensure required skills resolve (no “skill not found” errors).- If agent uses MCP, run `mcporter call <server>.bridge_health` first.## Common Patterns- **Health preflight:** run a health skill before sending user messages.- **Retry:** allow one automatic retry with a short delay for MCP send failures.- **Correlation IDs:** enforce a prefix per agent (`correlationIdPrefix`).## Troubleshooting| Issue | Fix || --- | --- || Agent missing in UI | File not saved in `agents/` or YAML invalid – re-open and validate. || Skill not found | Ensure the skill folder + tool YAML exist, and the name matches. || MCP errors | Bridge down or config mismatch – follow `skills/mcp/SKILL.md`. || Duplicate correlation IDs | Adjust `correlationIdPrefix` or generate UUIDs. |## Related Skills- `skills/agents-system` – bigger-picture maintenance- `skills/mcp` – MCP server operations- `skills/julia-bridge` – specific to the Julia bridge

Frequently asked questions

Agents