Skip to content
OpenAgentSkill
Browsing is open · Catalog captured 2026-11-21 · Automatic updates and account services are paused.
Search & research

pstack-omp

OMP adapter for poteto-mode. Maps canonical pstack roles and lifecycle protocols to current OMP task agents, batching, isolation, IRC follow-ups, and durable result resources.

dsebban · Static checks on file

Installation

Read the source and dependencies before running a command. Copying does not install or execute anything.

npx skills add dsebban/skills --skill pstack-omp

About this skill


name: pstack-omp description: "OMP adapter for poteto-mode. Maps canonical pstack roles and lifecycle protocols to current OMP task agents, batching, isolation, IRC follow-ups, and durable result resources."

pstack on OMP

poteto-mode is the sole router. It selects the playbook, canonical role, step order, and lifecycle protocol. This adapter translates those choices to OMP only. It never selects a playbook, repeats the playbook index, or changes a playbook gate.

While this skill is active, its role map is the specific pstack execution contract. Generic OMP instructions remain valid outside pstack work.

Canonical role map

Use these exact OMP agent names.

Canonical roleOMP agentContract
explorerscoutRead-only repository reconnaissance, trace reduction, narrow audits.
watcherscoutObserve one exact generation or external-state transition, then terminate.
plannerdesignerTechnical planning, architecture, decomposition, sequencing, and non-visual design candidates.
designerdesignerVisual, interaction, and product-design candidates.
reviewerreviewerIndependent code, protocol, behavioral, or security review.
researcherlibrarianSource-verified external library, framework, API, protocol, or version research.
synthesizerreviewerCross-report synthesis, adjudication, and advisory judgment over frozen evidence.
implementertaskBounded implementation or test changes with explicit write ownership.
ownertaskOne coupled multi-step implementation session retained through IRC follow-ups.
mechanicalsonicFully specified low-judgment edits. Ambiguity returns to the root.

The canonical OMP roster is closed:

scout
designer
reviewer
security-reviewer
librarian
task
sonic

A canonically routed task item's agent value must be one of those bundled names. Canonical role labels and model aliases are not agent names.

poteto-agent and comment-sicko are custom compatibility agents for direct named seams in imported skills. A direct compatibility call may use that custom agent name; ordinary canonical routing never does.

Planning distinction

The imported warning about a built-in planning subagent describes a source-host mechanism that bypassed the skill contract. This OMP version has no bundled plan agent. Canonical planner work uses designer with a technical-planning brief and does not pass source-host subagent fields.

Security review

Use security-reviewer for an independent security lane. Keep it read-only and separate from ordinary code review when both are required. A later reviewer session may synthesize frozen reports; it does not replace the primary security review.

Task contract

The root performs every task call. A child never calls task, starts another child, or asks the user directly. Put Do not call task or start subagents under FORBIDDEN in every child brief.

Children start without the parent conversation. Every prompt must stand alone. They receive the configured workspace, context files, skills, and approved plan resources supplied by OMP.

One task

When the flat schema is active:

{
  "name": "parser-overflow-worker",
  "agent": "task",
  "task": "GOAL\n...\n\nROLE\n...\n\nSCOPE\n...\n\nCONTEXT\n...\n\nACCEPTANCE\n...\n\nVERIFY\n...\n\nTIMEBOX\n...\n\nFORBIDDEN\n...\n\nREPORT\n...\n\nSTANDING\n...",
  "isolated": true
}

When batch mode is active, use a one-item tasks[] call instead of inventing a per-call batch switch.

Use isolated: true for a writer that needs a dedicated worktree. Read the returned isolation metadata and verify the patch, branch, or applied change OMP reports. Use a non-isolated session for a long-lived owner that must accept IRC follow-ups in the same workspace.

Batched panel

Use one batch call for independent participants:

{
  "context": "Shared immutable repository, base SHA, artifacts, constraints, and verification context.",
  "tasks": [
    {
      "name": "candidate-a",
      "agent": "designer",
      "task": "Standalone brief for architecture candidate A.",
      "isolated": false
    },
    {
      "name": "candidate-b",
      "agent": "designer",
      "task": "Standalone brief for architecture candidate B.",
      "isolated": false
    }
  ]
}

Each name is unique. context contains common immutable material. Each item still names its exact role, slice, acceptance criteria, verification, forbidden work, and report contract.

Start every participant in one batch before consuming any verdict. Separate writers with isolated workspaces or disjoint output paths. Freeze candidate artifacts before starting reviewers. Freeze reviewer reports before starting a separate synthesizer session.

Background completion and follow-ups

Background task results arrive through OMP's async result delivery. Record the returned job and agent identifiers. The preview may be truncated.

  • Read the full result from agent://<id>.
  • Read the session transcript from history://<id> when the report is incomplete, ambiguous, or suspicious.
  • Use hub with op: "jobs" to inspect jobs and op: "wait" with ids to wait for specific job IDs.
  • Use hub with op: "list" to inspect peers and op: "send" with to for one in-scope correction, answer, authorization, or next coupled phase.
  • Use hub with op: "cancel" and exact job ids for stale, superseded, or scope-breaching work.
  • A follow-up to a parked session revives the same session when the host reports that capability.
  • Never steer a reviewer toward a preferred conclusion.
  • Use only operations exposed by the live hub schema. Do not invent lifecycle operations.
  • Ignore duplicate terminal deliveries and reject stale artifact generations.

A task result is evidence, not completion. The root inspects the artifact and runs verification.

Brief shape

A di

Read the complete source ↗