Soften wave labels in public-facing copy

Map "Wave 1" and "Wave 2" prose to "first batch" and "follow-up batch"
across READMEs, docs, agent surfaces, and CLI help text. Internal
identifiers (`wave1`, `wave2`, `plan-wave2`, `planned_wave2_design`) are
preserved as stable artifact and CLI contracts. Tests updated alongside.

Author: jvogan

AGENTS.md

@@ -4,7 +4,7 @@ Public workspace for BioSymphony Ferm DoE.
 
 ## Mission
 
-Help users and their agents drive long-horizon, multi-agent bioprocess design campaigns from intake through Wave 2. The repo is a skill any agent harness can pick up: Symphony with Linear, Claude Code workers with Linear, Codex CLI, OpenAI Agents SDK, or a custom orchestrator, running on a laptop, in CI, or behind AWS Lambda or Modal. The agent and the human share one durable artifact (`campaign_manifest.json`), pause, resume, hand off, and converge on a fermentation campaign that is worth running.
+Help users and their agents drive long-horizon, multi-agent bioprocess design campaigns from intake through follow-up. The repo is a skill any agent harness can pick up: Symphony with Linear, Claude Code workers with Linear, Codex CLI, OpenAI Agents SDK, or a custom orchestrator, running on a laptop, in CI, or behind AWS Lambda or Modal. The agent and the human share one durable artifact (`campaign_manifest.json`), pause, resume, hand off, and converge on a fermentation campaign that is worth running.
 
 The repo ships:
 
@@ -53,7 +53,7 @@ Use synthetic or public-source examples only, label synthetic data clearly.
 
 ## Long-Agent Loop
 
-See `skills/biosymphony-ferm-doe/SKILL.md`. Summary: intake, readiness gate, factor framing, scale framing, DoE selection, run packet, Wave 2 review. Refuse on missing required blocks for declared profiles, missing assayed-response contracts, and public-safety violations. Warn on profile-advised gaps and DoE family minimum-runs shortfalls.
+See `skills/biosymphony-ferm-doe/SKILL.md`. Summary: intake, readiness gate, factor framing, scale framing, DoE selection, run packet, follow-up review. Refuse on missing required blocks for declared profiles, missing assayed-response contracts, and public-safety violations. Warn on profile-advised gaps and DoE family minimum-runs shortfalls.
 
 ## Key Pattern And Convention
 

CHANGELOG.md

@@ -75,7 +75,7 @@ catalog-demo:
 	PYTHONPATH=src $(PYTHON) -m biosymphony_ferm_doe.cli list-campaigns examples
 
 brief-demo:
-	PYTHONPATH=src $(PYTHON) -m biosymphony_ferm_doe.cli agent-brief examples/demo-pb-screening-public --goal "Plan a safe next-wave fermentation DOE campaign."
+	PYTHONPATH=src $(PYTHON) -m biosymphony_ferm_doe.cli agent-brief examples/demo-pb-screening-public --goal "Plan a safe follow-up fermentation DOE campaign."
 
 contract-check:
 	PYTHONPATH=src $(PYTHON) -m biosymphony_ferm_doe.cli validate-task-request templates/task_request.template.json
@@ -132,7 +132,7 @@ secret-scan-optional:
 release-check: test validate-all assay-power-demo contract-check tool-registry-check adaptive-backend-surface public-audit markdown-links
 	PYTHONPATH=src $(PYTHON) -m biosymphony_ferm_doe.cli inspect-campaign examples/demo-pb-screening-public --out /tmp/biosymphony-inspect-demo.json
 	PYTHONPATH=src $(PYTHON) -m biosymphony_ferm_doe.cli list-campaigns examples --out /tmp/biosymphony-campaign-catalog.json
-	PYTHONPATH=src $(PYTHON) -m biosymphony_ferm_doe.cli agent-brief examples/demo-pb-screening-public --goal "Plan a safe next-wave fermentation DOE campaign." --out /tmp/biosymphony-agent-brief.json --md-out /tmp/biosymphony-agent-brief.md
+	PYTHONPATH=src $(PYTHON) -m biosymphony_ferm_doe.cli agent-brief examples/demo-pb-screening-public --goal "Plan a safe follow-up fermentation DOE campaign." --out /tmp/biosymphony-agent-brief.json --md-out /tmp/biosymphony-agent-brief.md
 	PYTHONPATH=src $(PYTHON) -m biosymphony_ferm_doe.public_release \
 		README.md AGENTS.md pyproject.toml CHANGELOG.md NON_CLAIMS.md Makefile noxfile.py .gitignore \
 		BIOSAFETY.md SECURITY.md CONTRIBUTING.md CITATION.cff .github agents deploy scripts \

Makefile

@@ -10,10 +10,10 @@ Pre-experiment planning. It checks whether a fermentation campaign is measurable
 
 - Not a physical execution system. Not LIMS, not ELN, not robotics, not real-time bioreactor control, not scheduling.
 - Not a validated GxP batch-record system. Validators do not produce regulatory-grade evidence.
-- Not a replacement for JMP, Design-Expert, Modde, or a statistician. Wave 1 designs are emitted with a labeled statistical claim level (`exact` | `adapter_backed` | `approximate` | `heuristic`). Wave 1 analysis (`ferm-doe analyze`) provides stdlib OLS estimates with permutation p-values, bootstrap CIs, and lack-of-fit decomposition, labeled `wave1_analysis_planned`. The labeling is designed for statistician review before driving expensive Wave 2 runs, not as final regulatory-grade analysis.
+- Not a replacement for JMP, Design-Expert, Modde, or a statistician. first-batch designs are emitted with a labeled statistical claim level (`exact` | `adapter_backed` | `approximate` | `heuristic`). first-batch analysis (`ferm-doe analyze`) provides stdlib OLS estimates with permutation p-values, bootstrap CIs, and lack-of-fit decomposition, labeled `wave1_analysis_planned`. The labeling is designed for statistician review before driving expensive follow-up runs, not as final regulatory-grade analysis.
 - Not a physical-execution validation system for scale-up or scale-down. Scale-bridge artifacts in this repo describe planning, criteria, and qualification evidence the campaign would need; they do not validate that a small-scale model recapitulates a large-scale process unless executed runs with provenance and QC have been ingested.
-- Not a lab-execution cost, yield, or throughput model. `ferm-doe cost-rollup` multiplies operator-declared unit costs against artifact-derived counts (Wave 1 run count, sample count, sample volume, run duration) and emits a planning rollup labeled `cost_rollup_planning`; it does not estimate biological yields, robotics queueing, or staff time. Use it as a budgeting input, not as a quote. The five-tier cost stack documented in `docs/COST_MODEL_REALISM_CHECK.md` (simulator, bulk reagent, fully-loaded COGS, CMO benchmark, range) is the recommended honesty framing; a single $/mg number is never load-bearing.
-- Not an autonomous optimization system. Adaptive Wave 2 artifacts are planned next-step recommendations and learning records; they do not prove optimality, transferability, or execution readiness. BoFire and BoTorch adapter outputs are labeled `bayesian_optimization_planned`, not validated optimization.
+- Not a lab-execution cost, yield, or throughput model. `ferm-doe cost-rollup` multiplies operator-declared unit costs against artifact-derived counts (first-batch run count, sample count, sample volume, run duration) and emits a planning rollup labeled `cost_rollup_planning`; it does not estimate biological yields, robotics queueing, or staff time. Use it as a budgeting input, not as a quote. The five-tier cost stack documented in `docs/COST_MODEL_REALISM_CHECK.md` (simulator, bulk reagent, fully-loaded COGS, CMO benchmark, range) is the recommended honesty framing; a single $/mg number is never load-bearing.
+- Not an autonomous optimization system. Adaptive follow-up artifacts are planned next-step recommendations and learning records; they do not prove optimality, transferability, or execution readiness. BoFire and BoTorch adapter outputs are labeled `bayesian_optimization_planned`, not validated optimization.
 - Not a validated process simulator. Any simulator outputs (when present) carry a `simulator.fidelity_level` annotation on the manifest. A simulator number is one tier in the cost stack, not a substitute for the literature pressure-test at `templates/PARETO_LIT_CHECK.template.md` or for lab data.
 - Not a literature-discovery system. The cumulative-dossier pattern (per-corpus swarm, integrator, harvester) records what evidence rows the operator or a research worker provided, with provenance, confidence, and contradiction tracking. It does not perform unbounded web search inside the engine, and it does not promise topic coverage beyond what the operator supplies.
 
@@ -24,8 +24,8 @@ Pre-experiment planning. It checks whether a fermentation campaign is measurable
 - Profile-required blocks (e.g., scale_context for scale-bridge profiles).
 - DoE family minimum-runs guidance for declared families.
 - Per-axis readiness state (responses, factors, arms, scale_context, doe, decision_rules, evidence, feasibility).
-- Public Wave 2 planning checks after results are supplied: result-row QC/trust filtering, response-level assay-power policy, conservative next-action labels, arm-scoped negative memory, and learning-ledger handoff.
-- Stdlib OLS analysis of Wave 1 results: effect estimates, permutation p-values, residual-bootstrap CIs, lack-of-fit decomposition (when replicates exist), and an active-factor list to drive Wave 2 narrowing.
+- Public follow-up planning checks after results are supplied: result-row QC/trust filtering, response-level assay-power policy, conservative next-action labels, arm-scoped negative memory, and learning-ledger handoff.
+- Stdlib OLS analysis of first-batch results: effect estimates, permutation p-values, residual-bootstrap CIs, lack-of-fit decomposition (when replicates exist), and an active-factor list to drive follow-up narrowing.
 - Frictionless-compatible table contracts on CSV inputs (run ledger, evidence, equipment, reagent, design, results); stdlib by default, full Frictionless validation lights up with the `[contracts]` extra.
 - Curated 37-tool BO/DoE registry consistency at `docs/tool-registry.json` (adapter availability, route reasons, claim level).