phrazzld/dispatch
skills/dispatch/SKILL.md
Compose and launch roster-backed specialist lanes with prompt-native lane cards and receipts. Use when: "dispatch agents", "use subagents", "compose a team", "run provider lanes", "make lane cards". Trigger: /dispatch, /subagents, /lanes.
$ install --global
skillsauthnpx skillsauth add phrazzld/agent-skills dispatchInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: Jun 10, 2026, 6:23 AM22.4s7 files scanned
SKILL.md
- name:
- dispatch
- description:
- |
- cards and receipts. Use when:
- dispatch agents", "use subagents",
- Trigger:
- /dispatch, /subagents, /lanes.
- argument-hint:
- [objective|lane-card|provider target]
/dispatch
Global delegation, prompt-native. Native subagents are the default path; roster providers and sprites earn lanes on judgment, not quota. This skill keeps the implementation small: lane cards, native harness commands, receipts, and lead synthesis.
Route
| Need | Load |
|---|---|
| write or review lane cards | references/lane-cards.md |
| copy a lane card | templates/lane-card.md |
| track a multi-lane run | templates/run-card.md |
| choose providers/models | /harness-engineering models |
| run a lane on a remote sandbox | /sprites (sprite-lane runner) |
Use the existing dispatch-agent helper to launch configured providers. Use
lane_harness.v1 projection only when context hygiene matters; otherwise a
lane card is just the prompt packet.
Delegation Judgment
No provider quota. Native subagents by default; cross-model critics for
review of your own work; roster providers for bounded cards where the
provider is better, cheaper, or independent in a nameable way; sprites for
heavy, long-running, parallel, or isolation-needing lanes. See
harnesses/shared/AGENTS.md (Roster).
Local lane guidance: lane cards are outcome-shaped and big — end state, success criteria, verification affordances, boundaries, output shape, receipt expectations. The oracle field is load-bearing; the lane agent owns its own decomposition. Give each lane scoped context and record the lead's accepted or rejected evidence.
Contract
- The lead composes the team; Harness Kit does not rank providers or schedule fallback loops.
- A lane card is a natural-language contract, not a new runtime object.
- A provider receipt is evidence, not authority.
lane_harness.v1is an optional projection boundary for narrow lanes that should not inherit the full global skill catalog.- Failed providers produce typed evidence and return to the lead; replacement is explicit lead judgment.
Workflow
- Probe. Run the repo roster probe or equivalent smoke path.
- Plan lanes. Pick independent lanes: builder, critic, verifier, QA driver, performance critic, persona tester, product synthesizer, or a task-specific role.
- Write lane cards. Use
templates/lane-card.md. Keep each card small: one role, one objective, one boundary, one output shape. - Project only when needed. Add
--lane-harness <manifest>when the lane needs a focused visible skill set. - Dispatch in parallel. Launch independent providers with
dispatch-agent; do not wait serially unless a lane depends on another. - Synthesize. Accept, reject, or partially accept outputs. Name failures and waivers. The lead owns the final decision.
Output
## Dispatch Summary
- Objective:
- Providers used:
- Lane shape:
- Accepted evidence:
- Rejected evidence:
- Failures / waivers:
- Receipt ids:
- Residual risk:
Gotchas
- Do not create static subagent files for roles that a lane card can express.
- Do not generate or mutate global skill installs during a run.
- Do not count probes, auth failures, or wrapper failures as successful lanes.
- Do not add lanes for ceremony; each lane needs a distinct question, risk, artifact, or model/harness property. One good native subagent beats two ritual provider lanes.
- Do not use projection as a permission system; it is context hygiene.
Related Skills
phrazzld/roster
tools
VerifiedTrustedCommunity
Enumerates the peer AI agent CLIs installed on this machine (codex, claude, pi, opencode, cursor-agent, grok, agy, hermes, thinktank) and how to invoke each headlessly. A capability map, not a quota: useful for fresh-context adversarial review on a different model family, second opinions, competing attempts, and wide benches. Use when: "ask codex", "ask another model", "second opinion", "cross-model review", "what AI tools do I have", "other agents", "different model family", "adversarial critique from another provider". Trigger: /roster.
13SKILL.mdUpdated Jun 11, 2026
phrazzld/sprites
development
VerifiedTrustedCommunity
Run lane cards on Fly Sprites: remote, isolated, scale-to-zero sandboxes for heavy or parallel agent work. Golden-checkpoint provisioning so lanes start on a ready sprite with zero setup tokens. Use when: "run this on a sprite", "remote lane", "offload to a sandbox", "dispatch to sprites", "bake a sprite", "sprite fleet", heavy/long-running/parallel sub-agent work that should not run on this machine. Trigger: /sprites, /sprite-lane.
13SKILL.mdUpdated Jun 10, 2026
phrazzld/orient
tools
VerifiedTrustedCommunity
Fast session-start repository orientation from live local evidence. Use when: "orient yourself", "start of session", "new session", "where are we", "catch me up before acting", "what should I do next", after compaction, after switching worktrees, or before choosing a Harness Kit workflow. Trigger: /orient, /ground, /session-start.
13SKILL.mdUpdated Jun 7, 2026
phrazzld/debrief
development
VerifiedTrustedCommunity
Lightweight evidence-backed retro and catch-up reports for a current repo, branch, PR, backlog slice, or recent agent session. Use when the user asks for a debrief, catch me up, what changed, why it matters, product implications, end-user implications, developer experience implications, current app state, backlog state, workspace state, alternatives considered, or context rebuild after losing the thread. Trigger: /debrief.
13SKILL.mdUpdated Jun 5, 2026