Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

Agent Native (parallel agent dispatch)

Dispatch independent tasks to parallel workers or subagents without write-scope collisions, then observe and report runtime evidence per packet.

AutomationAdvanced★ 448⑂ 41AI score 7/10Last updated: Oct 11, 2026

What it does

  • Validates the entire batch before launch: every packet must declare its executor, packet identity, transitive effects and a canonical workspace-relative write scope. If any packet is invalid, nothing launches.
  • Detects write-scope collisions by resolving symlinks, normalizing paths and comparing case-insensitively, so aliases can't hide overlaps.
  • Refuses "prompt-send optimism": delivery proves transport only; engagement must be shown by session output, tool activity or changed files.
  • Captures worker state before any nudge, replacement or restart, so the reason for a stall isn't destroyed.
  • Defines role authority (Orchestrator, Implementer, Validator, Scribe) plus bounded Reader/Writer delegations with explicit limits.
  • Returns a per-packet record: packet_id, executor, observed_session, engagement, terminal_status, result, gaps.

Who it's for

  • Engineers running multiple Claude/Codex sessions concurrently who have hit file clobbering or duplicate retries.
  • Users of the AgentOps plugin ecosystem (orchestrate, codex-exec, claude-exec).
  • Teams needing an auditable evidence trail for multi-agent pipelines.

Examples

  1. "Refactor module A and module B at the same time" → canonical write scopes are compared, proven disjoint, then both launch together.
  2. A worker goes quiet for ten minutes → instead of restarting, capture its observable state first so the stall cause is preserved.
  3. An executor errors out → no automatic retry; the error is recorded verbatim as that packet's result and repair stays with the caller.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/boshu2/agentops/HEAD/plugin/skills/agent-native/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the plugin/skills/agent-native folder from the GitHub repo boshu2/agentops into my ~/.claude/skills/boshu2-agent-native-3/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/boshu2/agentops.git && mkdir -p ~/.claude/skills && cp -r agentops/plugin/skills/agent-native ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal.
  2. Clone the repo: git clone https://github.com/boshu2/agentops.git
  3. Create the skills directory: mkdir -p ~/.claude/skills
  4. Copy the skill: cp -r agentops/plugin/skills/agent-native ~/.claude/skills/
  5. This skill links to sibling skills (orchestrate, codex-exec, claude-exec), so copying the whole set is recommended: cp -r agentops/plugin/skills/* ~/.claude/skills/
  6. Restart Claude Code and test with a request like "run these agents in parallel" to confirm it triggers.
  7. Note: the reference implementation scripts/swarm/dispatch_once.py is not bundled with installed skills — run it from the cloned AgentOps source checkout if needed.
View source on GitHub ↗License: Apache-2.0