AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Agent Dispatch

skill-conn-castle-agent-layer-agent-dispatch · by conn-castle

Use `al dispatch` only when the user names an external dispatch target or another skill explicitly requires dispatch. Treat generic subagent, second-agent, and fresh-context requests as built-in subagent work.

↗ Website ↗ Repository
— No reviews yet
0 installs
18 views
0.0% view→install

Install

$ agentstack add skill-conn-castle-agent-layer-agent-dispatch

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • ✓ Prompt-injection patterns
  • ✓ Secret / credential exfiltration
  • ✓ Dangerous shell & filesystem operations
  • ✓ Untrusted network calls
  • ✓ Known-malicious package signatures

What it can access

  • ✓ Network access No
  • ✓ Filesystem access No
  • ✓ Shell / process execution No
  • ✓ Environment & secrets No
  • ✓ Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-conn-castle-agent-layer-agent-dispatch)

Reliability & compatibility

✓ Security review passed
0 installs to date
— no reviews yet
● 1mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Agent Dispatch? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

al dispatch

Required artifacts

  • No artifact is required for ordinary al dispatch runs.
  • If preparing prompt files or saving target output, put agent-only files under

.agent-layer/tmp/.

  • Use .agent-layer/tmp/agent-dispatch...md or

.agent-layer/tmp/agent-dispatch...json for scratch prompt, output, or options files.

  • Report every artifact path created.

Global constraints

  • Run al dispatch --help before the first al dispatch command in a session.
  • Run relevant subcommand help before using non-obvious subcommands or flags.
  • If al, a target CLI, authentication, source skill, or generated skill

projection is missing, stop and report the missing requirement. Do not install, authenticate, sync, or change configuration unless the user asked for setup or repair.

Workflow

  1. Verify any requested skill exists in the current project's Agent Layer skill

source. Do not invent a skill name.

  1. Send the focused prompt to one target. Omit unrelated conversation

history and information that could bias the target.

  1. If a prompt file is needed, write it under .agent-layer/tmp/.
  2. Run al dispatch and wait for the original process to finish, even if

output is quiet or it appears stalled. Do not poll. Wait until it finishes to inspect dispatch artifacts.

Guardrails

  • Do not use al dispatch for ordinary local shell work, tests, web retrieval,

browser automation, or API-only tasks.

  • Do not use al dispatch to bypass the caller's restrictions, approvals, or

sandbox expectations.

  • Do not retry failed target launches by guessing flags, targets, models, or

reasoning-effort values. Re-check help and options, then report the mismatch if it remains unresolved.

Definition of done

  • Current target metadata was inspected before any real run.
  • Target selection and skill delivery if used were intentional and reported.
  • Prompt and output artifacts were stored only under .agent-layer/tmp/ and

every created artifact path was reported.

  • Errors were reported explicitly.
  • Any side effects from target work were inspected and verified through the

current task's normal checks.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Write a review

Versions

  • v0.1.0 Imported from the upstream source.

More in Agent Skills