babyworm/rtl-agent-team/rtl-p4s-unit-test

<Purpose> Write and run unit tests that verify each RTL module implements its microarchitecture specification correctly. Compares RTL output against the C reference model for functional correctness. Goes beyond Tier 1 smoke testing (rtl-p4-implement Wave 6) to exercise key features at uarch level: FSM state transitions, pipeline stage behavior, data transformation accuracy.

Testing Tier Context:

Tier 1: Smoke Test     — connectivity, R/W, basic ops (rtl-p4-implement Wave 6)
Tier 2: Unit Test      — reference comparison, uarch features (THIS SKILL) ←
Tier 3: Module Regr.   — cocotb multi-seed (rtl-p5s-func-verify)
Tier 4: Integration    — cross-module, end-to-end (rtl-p5s-integration-test)

Outputs: sim/{module}/tb_{module}.sv testbench files + sim/{module}/{module}_unit_results.json. </Purpose>

<Use_When>

• Phase 4 RTL is lint-clean AND Tier 1 smoke test passed (rtl-p4-implement Wave 6)
• Each module needs uarch-level functional verification with reference comparison
• A new module's key features need targeted testing beyond connectivity
• A bug fix needs regression test verifying the fix against reference model

</Use_When>

<Do_Not_Use_When>

• Only basic connectivity/R/W verification needed (covered by rtl-p4-implement Wave 6 smoke — Tier 1)
• Full multi-seed regression needed (use rtl-p5s-func-verify — Tier 3)
• Integration/cross-module testing (use rtl-p5s-integration-test — Tier 4)
• Formal verification preferred (use rtl-p5s-sva-check instead)

</Do_Not_Use_When>

<Why_This_Exists> Tier 1 (smoke) only verifies connectivity and basic I/O. Tier 2 unit tests verify that each module implements its uarch spec — FSM states, pipeline stages, data transformations — using the C reference model as the golden oracle. This catches behavioral bugs that smoke tests miss, while remaining faster and more targeted than full regression (Tier 3). </Why_This_Exists>

<Delegation> The orchestrator writes testbenches per module (parallel via testbench-dev), selects reference comparison mode (DPI-C or file-based), runs simulations, and handles failure triage with waveform analysis.

All coding conventions, reference mode rules, escalation criteria, and result schemas are defined in the rtl-p4s-unit-test-policy skill (loaded via the orchestrator's skills: field).

Execution

Task(subagent_type="rtl-agent-team:p4s-unit-test-orchestrator", prompt="Execute Tier 2 unit testing. User input: $ARGUMENTS")

Do not perform any work directly. The orchestrator agent manages testbench generation, simulation, and failure triage.

Output

• sim/{module}/tb_{module}.sv — Tier 2 unit test testbenches
• sim/{module}/{module}_unit_results.json — results with ref_mismatches=0, coverage thresholds met, and req_ids populated

</Delegation>