Run Parity Test J

run-parity-test-j — Entry Skill

> Entry point for the parity-test-j pipeline. Orchestrates the sequential native-subagent dispatch of the analyzer, reviewer, and reporter agents on Tier 1b (OpenCode v1.15.10). Each step runs as a distinct subagent process — no inline execution, no Task() calls.

This skill initializes the pipeline run, resolves the scope root and run ID, surfaces the required Tier 1b degradation warning, dispatches the analyzer subagent, dispatches the reviewer subagent (structural isolation: plan + disallowedTools), dispatches the reporter subagent (only on approved verdict), and applies the cleanup contract (C20) on completion. Model is declared per-agent in frontmatter (opencode/big-pickle); the entry skill itself uses disable-model-invocation: true.

Platform Context

• Tier: tier_1b (OpenCode v1.15.10)
• Dispatch mechanism: native_subagent — DISPATCH(mode="subagent", agent="{name}", context={...}). NOT Task(). NOT inline.
• modelfieldformat: provider_prefixed — model: opencode/big-pickle declared in each agent's YAML frontmatter.
• Reviewer isolation: structural — reviewer runs with permissionMode: plan + disallowedTools: Write, Edit, Bash.
• Degradation warnings: 1 active (see PHASE 0 below).

DEGRADATION WARNINGS (Tier 1b — Active)

The following warning from the tier_1b platform profile MUST be surfaced before execution:

1. Parallel fan-out (Pattern 2) degrades to sequential on OpenCode.

Workflow

PHASE 0: PREFLIGHT — DEGRADATION WARNING AND INITIALIZATION

Step 0.1 — Surface Tier 1b degradation warning (MANDATORY before any execution):

Surface the following warning to the user verbatim before any other action:

> [Tier 1b Degradation Warning 1 of 1] Parallel fan-out (Pattern 2) degrades to sequential on OpenCode.

Step 0.2 — Resolve runtime context:

1. Resolve {ROOT} via sk-pipeline-paths (scope root resolved from the active tier profile at runtime — NEVER hardcode .opencode/, .superpipelines/, .claude/, or any platform path).
2. Resolve {runId} = ISO-8601 compact timestamp (e.g., 20260526T143000Z).
3. Create temp directory: {ROOT}/superpipelines/temp/parity-test-j/{runId}/.
4. Create output/ directory if absent: {ROOT}/output/.

Step 0.3 — Collect inputs:

If the user has not supplied the path to the input CSV file, ask for it now. Record as {INPUT_PATH}.

Step 0.4 — Initialize pipeline-state.json:

Write pipeline-state.json to {ROOT}/superpipelines/temp/parity-test-j/{runId}/pipeline-state.json:

{
  "pipeline_id": "parity-test-j",
  "run_id": "{runId}",
  "started_at": "{iso8601}",
  "plugin_version": "2.0.0",
  "pattern": "1",
  "status": "running",
  "current_phase": 0,
  "metadata": {
    "source_tier": "tier_1b",
    "runtime_tier": "tier_1b",
    "model_field_format": "provider_prefixed",
    "resolved_models": {
      "analyzer": "opencode/big-pickle",
      "reviewer": "opencode/big-pickle",
      "reporter": "opencode/big-pickle"
    }
  },
  "phases": [
    {
      "index": 0,
      "step_id": "analyzer",
      "name": "analyze",
      "status": "pending",
      "agent": "agents/superpipelines/parity-test-j/analyzer.md",
      "outputs": [],
      "error": null
    },
    {
      "index": 1,
      "step_id": "reviewer",
      "name": "review",
      "status": "pending",
      "agent": "agents/superpipelines/parity-test-j/reviewer.md",
      "reviewer_isolation": "structural",
      "outputs": [],
      "error": null
    },
    {
      "index": 2,
      "step_id": "reporter",
      "name": "render",
      "status": "pending",
      "agent": "agents/superpipelines/parity-test-j/reporter.md",
      "outputs": [],
      "error": null
    }
  ]
}

PHASE 1: DISPATCH ANALYZER (native subagent)

Dispatch the analyzer as a native subagent:

DISPATCH(
  mode="subagent",
  agent="analyzer",
  context={
    "input_path": "{INPUT_PATH}",
    "findings_output_path": "{ROOT}/superpipelines/temp/parity-test-j/{runId}/findings.json",
    "state_path": "{ROOT}/superpipelines/temp/parity-test-j/{runId}/pipeline-state.json",
    "run_id": "{runId}",
    "root": "{ROOT}"
  }
)

Wait for terminal status from the analyzer subagent:

• DONE → update pipeline-state.json phases[0].status = "completed"; advance to Phase 2.
• DONE_WITH_CONCERNS → update phases[0].status = "completedwithconcerns"; surface concerns to user; advance to Phase 2.
• NEEDS_CONTEXT → update phases[0].status = "blocked"; update top-level status = "blocked"; surface message to user; GO TO CLEANUP (preserve).
• BLOCKED → update phases[0].status = "blocked"; update top-level status = "blocked"; surface message to user; GO TO CLEANUP (preserve).

PHASE 2: DISPATCH REVIEWER (native subagent — structural isolation)

Dispatch the reviewer as a native subagent. The reviewer has permissionMode: plan and disallowedTools: Write, Edit, Bash — it cannot write files. Capture the REVIEWER VERDICT block from its terminal output text.

DISPATCH(
  mode="subagent",
  agent="reviewer",
  context={
    "findings_path": "{ROOT}/superpipelines/temp/parity-test-j/{runId}/findings.json",
    "state_path": "{ROOT}/superpipelines/temp/parity-test-j/{runId}/pipeline-state.json",
    "run_id": "{runId}",
    "root": "{ROOT}"
  }
)

After the reviewer subagent completes:

1. Extract the REVIEWER VERDICT block from the reviewer's terminal output text. Record as {REVIEWER_OUTPUT}.
2. Parse verdict from the block: approved, approved_with_concerns, or rejected.
3. Update pipeline-state.json:

• phases[1].outputs → [{ "verdict": "{verdict}", "reviewer_notes": "{REVIEWER_OUTPUT}" }]

Wait for terminal status:

• DONE (verdict: approved) → update phases[1].status = "completed"; advance to Phase 3.
• DONE_WITH_CONCERNS (verdict: approvedwithconcerns) → update phases[1].status = "completedwithconcerns"; surface concerns; advance to Phase 3.
• BLOCKED (verdict: rejected or validation failure) → update phases[1].status = "blocked"; update top-level status = "blocked"; surface rejection message; GO TO CLEANUP (preserve). DO NOT dispatch reporter.
• NEEDS_CONTEXT → update phases[1].status = "blocked"; update top-level status = "blocked"; GO TO CLEANUP (preserve).

PHASE 3: DISPATCH REPORTER (native subagent)

Only dispatched if reviewer verdict is approved or approved_with_concerns.

DISPATCH(
  mode="subagent",
  agent="reporter",
  context={
    "findings_path": "{ROOT}/superpipelines/temp/parity-test-j/{runId}/findings.json",
    "reviewer_verdict": "{verdict}",
    "reviewer_notes": "{REVIEWER_OUTPUT}",
    "report_output_path": "{ROOT}/output/parity-test-j-report.md",
    "state_path": "{ROOT}/superpipelines/temp/parity-test-j/{runId}/pipeline-state.json",
    "run_id": "{runId}",
    "root": "{ROOT}"
  }
)

Wait for terminal status from the reporter subagent:

• DONE → update pipeline-state.json phases[2].status = "completed"; advance to Phase 4.
• DONE_WITH_CONCERNS → update phases[2].status = "completedwithconcerns"; surface concerns to user; advance to Phase 4.
• NEEDS_CONTEXT → update phases[2].status = "blocked"; update top-level status = "blocked"; GO TO CLEANUP (preserve).
• BLOCKED → update phases[2].status = "blocked"; update top-level status = "blocked"; surface message; GO TO CLEANUP (preserve).

PHASE 4: FINALIZE

1. Update pipeline-state.json:

• status: "completed" (or "completed_with_concerns" if any phase had concerns)
• completed_at: ISO-8601 timestamp

1. Cleanup contract (C20):

• On DONE or DONE_WITH_CONCERNS: write status: completed to pipeline-state.json, then delete the temp directory {ROOT}/superpipelines/temp/parity-test-j/{runId}/.
• On BLOCKED or NEEDS_CONTEXT: preserve the temp directory (do NOT delete); it holds resume state for diagnosis.

1. Confirm to the user:

> Pipeline parity-test-j completed. Data quality document written to: {ROOT}/output/parity-test-j-report.md

1. Emit terminal status: DONE (or DONE_WITH_CONCERNS / BLOCKED / NEEDS_CONTEXT as applicable).

• NEVER call Task() — Tier 1b has task_primitive: false. Use DISPATCH(mode="subagent", ...) exclusively.
• NEVER execute steps inline — each step MUST run as a distinct native subagent process.
• NEVER hardcode platform paths (.opencode/, .superpipelines/, .claude/, etc.) — always use {ROOT} resolved via sk-pipeline-paths.
• NEVER pass file contents in dispatch context — pass file paths only (anti-pattern #3 Context Dumping).
• NEVER dispatch the reporter if the reviewer verdict is rejected or the reviewer emits BLOCKED.
• ALWAYS surface the Tier 1b degradation warning before any execution begins (Phase 0.1 is mandatory).
• ALWAYS capture the REVIEWER VERDICT block from the reviewer's terminal output before advancing.
• ALWAYS update pipeline-state.json after each phase completes.
• NEVER advance past a BLOCKED or NEEDS_CONTEXT status without human input.
• ALWAYS apply the C20 cleanup contract: delete temp dir on DONE/DONEWITHCONCERNS, preserve on BLOCKED/NEEDS_CONTEXT.
• Emit exactly one terminal status at the end of this skill's own execution: DONE / DONEWITHCONCERNS / NEEDS_CONTEXT / BLOCKED.

Source & license

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

• Author: gustavo-meilus
• Source: gustavo-meilus/superpipelines
• License: MIT

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