Run Parity Test H

run-parity-test-h — Entry Skill

> Entry point for the parity-test-h pipeline. Orchestrates the sequential inline execution of the validator, reviewer, and reporter protocol steps on Tier 2 (Cursor/Windsurf/Cline). All steps execute in this skill's session — no subagents, no Task() calls.

This skill initializes the pipeline run, resolves the scope root and run ID, surfaces all required Tier 2 degradation warnings, executes the validator, reviewer, and reporter protocol skills inline (same session, sequential), and applies the cleanup contract (C20) on completion. Model is inherit throughout — the host IDE owns model selection.

Platform Context

• Tier: tier_2 (Cursor/Windsurf/Cline)
• Dispatch mechanism: inline — no Task() primitive, no subagent processes. All three steps execute within this skill's session.
• modelfieldformat: omit — no model: field is emitted for any step. The host IDE owns model selection.
• Reviewer isolation: convention — no structural isolation. C19 self-skepticism preamble required in the reviewer protocol.
• Degradation warnings: 3 active (see PHASE 0 below).

Workflow

PHASE 0: PREFLIGHT — DEGRADATION WARNINGS AND INITIALIZATION

Step 0.1 — Surface Tier 2 degradation warnings (MANDATORY before any execution):

Surface ALL of the following warnings to the user verbatim before any other action:

> [Tier 2 Degradation Warning 1 of 3] Reviewer isolation impossible on this platform: writer and reviewer are the same agent in the same context. Review steps execute the reviewer protocol but cannot provide assumption-blindness defense or context-bleed isolation. Treat review output as a self-check, not verification.

> [Tier 2 Degradation Warning 2 of 3] Parallel fan-out (Pattern 2) requires worktrees and is unavailable on this platform (worktrees: false).

> [Tier 2 Degradation Warning 3 of 3] Model selection is owned by the host IDE; per-step model assignment is not emitted.

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 .superpipelines/, .cursor/, .claude/, or any platform path).
2. Resolve {runId} = ISO-8601 compact timestamp (e.g., 20260526T143000Z).
3. Create temp directory: {ROOT}/superpipelines/temp/parity-test-h/{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 YAML config 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-h/{runId}/pipeline-state.json:

{
  "pipeline_id": "parity-test-h",
  "run_id": "{runId}",
  "started_at": "{iso8601}",
  "plugin_version": "2.0.0",
  "pattern": "1",
  "status": "running",
  "current_phase": 0,
  "metadata": {
    "source_tier": "tier_2",
    "runtime_tier": "tier_2",
    "model_field_format": "omit",
    "resolved_models": {}
  },
  "phases": [
    {
      "index": 0,
      "step_id": "validator",
      "name": "validate",
      "status": "pending",
      "agent": null,
      "outputs": [],
      "error": null
    },
    {
      "index": 1,
      "step_id": "reviewer",
      "name": "review",
      "status": "pending",
      "agent": null,
      "outputs": [],
      "error": null
    },
    {
      "index": 2,
      "step_id": "reporter",
      "name": "report",
      "status": "pending",
      "agent": null,
      "outputs": [],
      "error": null
    }
  ]
}

Note: agent is null in all phases — Tier 2 has no agent files.

PHASE 1: EXECUTE VALIDATOR (inline)

Load protocol skill into context:

Protocol: {ROOT}/skills/superpipelines/parity-test-h/validator-protocol/SKILL.md

Execute the validator protocol inline (same session) with the following context:

input_path: {INPUT_PATH}
findings_output_path: {ROOT}/superpipelines/temp/parity-test-h/{runId}/validator-findings.json
state_path: {ROOT}/superpipelines/temp/parity-test-h/{runId}/pipeline-state.json
run_id: {runId}
root: {ROOT}

Wait for terminal status from the validator protocol:

• 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: EXECUTE REVIEWER (inline)

Load protocol skill into context:

Protocol: {ROOT}/skills/superpipelines/parity-test-h/reviewer-protocol/SKILL.md

Execute the reviewer protocol inline (same session) with the following context:

input_path: {INPUT_PATH}
findings_path: {ROOT}/superpipelines/temp/parity-test-h/{runId}/validator-findings.json
reviewed_findings_output_path: {ROOT}/superpipelines/temp/parity-test-h/{runId}/reviewed-findings.json
state_path: {ROOT}/superpipelines/temp/parity-test-h/{runId}/pipeline-state.json
run_id: {runId}
root: {ROOT}

Note: input_path is passed to the reviewer so it can independently re-read the source YAML per the C19 assumption-blindness requirement. The reviewer MUST re-derive findings from the raw YAML, not from the validator's summary.

Wait for terminal status from the reviewer protocol:

• DONE → update pipeline-state.json phases[1].status = "completed"; advance to Phase 3.
• DONE_WITH_CONCERNS → update phases[1].status = "completedwithconcerns"; surface concerns to user; advance to Phase 3.
• NEEDS_CONTEXT → update phases[1].status = "blocked"; update top-level status = "blocked"; surface message to user; GO TO CLEANUP (preserve).
• BLOCKED → update phases[1].status = "blocked"; update top-level status = "blocked"; surface message to user; GO TO CLEANUP (preserve).

PHASE 3: EXECUTE REPORTER (inline)

Load protocol skill into context:

Protocol: {ROOT}/skills/superpipelines/parity-test-h/reporter-protocol/SKILL.md

Execute the reporter protocol inline (same session) with the following context:

reviewed_findings_path: {ROOT}/superpipelines/temp/parity-test-h/{runId}/reviewed-findings.json
output_path: {ROOT}/output/parity-test-h-validation-report.md
state_path: {ROOT}/superpipelines/temp/parity-test-h/{runId}/pipeline-state.json
run_id: {runId}
root: {ROOT}

Wait for terminal status from the reporter protocol:

• DONE → update 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"; surface message to user; GO TO CLEANUP (preserve).
• BLOCKED → update phases[2].status = "blocked"; update top-level status = "blocked"; surface message to user; 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-h/{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-h completed. Validation report written to: {ROOT}/output/parity-test-h-validation-report.md

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

• NEVER call Task() — Tier 2 has task_primitive: false and subagents: false. All execution is inline.
• NEVER emit a model: field for any step — model_field_format: omit; host IDE owns model selection.
• NEVER hardcode platform paths (.superpipelines/, .cursor/, .claude/, etc.) — always use {ROOT} resolved via sk-pipeline-paths.
• NEVER pass file contents in execution context — pass file paths only (anti-pattern #3 Context Dumping).
• ALWAYS surface all 3 Tier 2 degradation warnings before any execution begins (Phase 0.1 is mandatory).
• ALWAYS pass input_path to the reviewer so it can independently re-read the source YAML (C19 assumption-blindness).
• 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.