feat(thinking): Round 69 prompt contract standardization

Six-lens panel review agreed to add Purpose/Scope/Acceptance/Falsifiability
contracts to all 01-thinking prompts, map lenses to frozen workflow skills,
and gate quality with test_thinking_prompts_eval_harness.py (80%+ eval_harness).

Also: Decision Index fix, QUALITY_GATES 200+ pytest floor, panel Round 69 record.

Author: johnteee

reflective-prompt-library/01-thinking/counterargument.md

@@ -2,6 +2,31 @@
 
 Use this to prevent excessive optimism, overengineering, or AI flattery.
 
+## Purpose
+
+Stress-test optimism, overengineering, and AI flattery before committing resources. Primary workflow surfaces: `reflective-review` and `reflective-minimality`; escalate to `reflective-risk` when trust-boundary or blast-radius signals appear.
+
+## Scope
+
+- In scope: attacking a proposed plan, architecture, or feature scope for fatal flaws and cheaper alternatives.
+- Out of scope: polite encouragement, implementation work, or net-new research without a plan to attack.
+
+## Acceptance Criteria
+
+- Fatal flaws separated from major and minor concerns.
+- At least one lower-cost alternative named.
+- Minimum viable version stated.
+- Kill criteria defined (when to stop).
+
+## Falsifiability
+
+List kill criteria: what evidence or outcome means the proposal should be abandoned rather than iterated.
+
+## Human Review
+
+Escalate to `reflective-risk` with an explicit Human Review gate when the work implies irreversible or high-blast-radius action.
+
+
 ```markdown
 請扮演最嚴格但公平的反方架構師，攻擊我的方案。
 
@@ -30,4 +55,3 @@ Use this to prevent excessive optimism, overengineering, or AI flattery.
 - Minimum viable version：最小可行版本
 - Kill criteria：什麼情況應該停止
 ```
-

reflective-prompt-library/01-thinking/critical-thinking-check.md

@@ -2,6 +2,31 @@
 
 Use this to review articles, strategies, product ideas, technical choices, or AI-generated claims.
 
+## Purpose
+
+Audit claims, assumptions, and evidence quality before accepting a plan, article, or AI output. Primary workflow surface: `reflective-review`; pair with `reflective-research` when external sources are required.
+
+## Scope
+
+- In scope: articles, strategies, product ideas, technical choices, AI-generated claims, diff/plan review prep.
+- Out of scope: writing first-draft specs (`reflective-spec-plan`), repository implementation (`reflective-implement`), formal-environment gating (`reflective-risk`).
+
+## Acceptance Criteria
+
+- Major and minor claims listed separately.
+- Assumption audit distinguishes explicit vs implicit risks.
+- Strongest counterargument stated (not a strawman).
+- Each conclusion tagged: supported, plausible-but-unproven, or unsupported.
+
+## Falsifiability
+
+Name at least one observation, dataset, or experiment that would overturn the strongest supported conclusion.
+
+## Human Review
+
+Escalate to `reflective-risk` with an explicit Human Review gate when the work implies irreversible or high-blast-radius action.
+
+
 ```markdown
 請用 Critical Thinking Review 嚴格審查以下內容：
 
@@ -61,4 +86,3 @@ Use this to review articles, strategies, product ideas, technical choices, or AI
 - 需要更多證據部分
 - 不建議採用部分
 ```
-

reflective-prompt-library/01-thinking/falsifiability.md

@@ -2,6 +2,31 @@
 
 Use this for scientific thinking, product experiments, AI workflows, and business models.
 
+## Purpose
+
+Turn vague ideas into testable hypotheses before experiments or delivery commitments. Primary workflow surfaces: `reflective-brief` for early framing; `reflective-spec-plan` when the hypothesis becomes a test plan.
+
+## Scope
+
+- In scope: scientific thinking, product experiments, AI workflow checks, business model bets.
+- Out of scope: full implementation, formal-environment rollout, or external source surveys without a stated hypothesis.
+
+## Acceptance Criteria
+
+- Hypothesis is binary and observable.
+- Prediction and falsifier are distinct.
+- Minimum test, metrics, and timebox are concrete (no "feels better").
+- Decision rule maps outcomes to next action.
+
+## Falsifiability
+
+Every output must include an explicit falsifier and a minimum test; if none can be stated, the idea is not ready for execution.
+
+## Human Review
+
+Escalate to `reflective-risk` with an explicit Human Review gate when the work implies irreversible or high-blast-radius action.
+
+
 ```markdown
 請把以下想法改寫成可證偽假說。
 
@@ -22,4 +47,3 @@ Use this for scientific thinking, product experiments, AI workflows, and busines
 
 請避免空泛指標，例如「感覺更好」。請盡量轉成可觀察、可比較、可重複的證據。
 ```
-

reflective-prompt-library/01-thinking/socratic-reviewer.md

@@ -2,6 +2,26 @@
 
 Suitable for requirements interviews, life decisions, product direction, business models, technical selection, learning strategy, and research question definition.
 
+## Purpose
+
+Clarify the real question before choosing a direction. Primary workflow surfaces: `reflective-brief` for goal and assumption clarification; escalate to `reflective-spec-plan` when scope is clear enough to plan.
+
+## Scope
+
+- In scope: requirements interviews, product direction, technical selection, learning strategy, research question definition.
+- Out of scope: code implementation (`reflective-implement`), production risk gating (`reflective-risk`), source-backed external research (`reflective-research`).
+
+## Acceptance Criteria
+
+- Core problem stated in one sentence.
+- At least one high-risk assumption named.
+- Minimal falsifiable test or observation defined.
+- Explicit defer list when scope is still open.
+
+## Falsifiability
+
+If the session cannot name what evidence would prove the current framing wrong, stop and return to Clarify instead of recommending action.
+
 ```markdown
 你是 Socratic Questioner。你的目標不是立刻給答案，而是幫我逼近真正問題。
 
@@ -52,4 +72,3 @@ Suitable for requirements interviews, life decisions, product direction, busines
 - 下一步
 - 暫時不該做的事
 ```
-

reflective-prompt-library/01-thinking/why-what-how-done.md

@@ -2,6 +2,26 @@
 
 Use this as the core gate prompt before committing to a direction.
 
+## Purpose
+
+Gate a task through Why / What / How / Done before choosing strictness or workflow depth. Primary workflow surfaces: `reflective-brief` then `reflective-dispatch` for orchestration level selection.
+
+## Scope
+
+- In scope: pre-commitment checks on goal, scope, method, and completion evidence for a single task or feature.
+- Out of scope: post-hoc code review (`reflective-review`), handoff retros (`reflective-handoff-retro`), or detailed test implementation.
+
+## Acceptance Criteria
+
+- Why gate answers cost of doing nothing and doing wrong.
+- What gate lists scope in/out, outputs, and acceptance criteria.
+- How gate names minimal method, risks, and human review triggers.
+- Done gate ties completion to observable evidence.
+
+## Falsifiability
+
+Done gate must name evidence that would prove the task should not proceed or should be rolled back.
+
 ```markdown
 請把任務通過 Why / What / How / Done 四層檢查。
 
@@ -52,4 +72,3 @@ Use this as the core gate prompt before committing to a direction.
 - 應該做到哪個層級
 - Prompting 即可，還是需要 Agent / Workflow / 程式
 ```
-

reflective-prompt-library/PROJECT_KNOWLEDGE.md

@@ -72,10 +72,11 @@ deferred promotions are recurrence-gated — see [panel backlog](plans/multi-age
 
 ## Decision Index
 
-- 2026-06-25 Post-Round 68 holdout expansion — ROUTE-003 minimality trap + thin adversarial group refresh (15 groups / 53 phrases) → [QUALITY_GATES_SUMMARY.md](plans/QUALITY_GATES_SUMMARY.md)
 > Pointers to the causal trail — plans, reflections, tests, commits. Detail is
 > not duplicated here; this is a map, not an archive.
 
+- 2026-06-25 Round 69 panel — standardize `01-thinking/` prompt contracts (Purpose/Scope/Acceptance/Falsifiability) + eval_harness anti-drift → [record](plans/multi-agent-panel-consensus-2026-06-25.md)
+- 2026-06-25 Post-Round 68 holdout expansion — ROUTE-003 minimality trap + thin adversarial group refresh (15 groups / 53 phrases) → [QUALITY_GATES_SUMMARY.md](plans/QUALITY_GATES_SUMMARY.md)
 - 2026-06-11 Add `runtime-trust-boundary` as a lens, not a ninth skill → [record](plans/project-adjustment-reflection-2026-06-11.md)
 - 2026-06-17 Adopt a minimal project-judgment contract instead of importing Knowie → [initial record](plans/knowie-project-knowledge-reflection-2026-06-17.md)
 - 2026-06-17 Refine the contract from non-normative to non-authoritative, connect promotion, and publish a reusable scaffold → [decision](plans/project-knowledge-authority-promotion-decision-2026-06-17.md)

reflective-prompt-library/plans/QUALITY_GATES_SUMMARY.md

@@ -314,7 +314,7 @@ ROUTE-002 measures unseen phrasing separately from ROUTE-001. Round 7 (2026-06-2
 2. **ROUTE-001/002/003 in CI** — 128 + 102 + 53 paraphrases at 100% consistency (seeded fixtures); `validate_route_fixture.py` gates minimum coverage
 3. **Governance validators** — links, lint, governance metadata, PROJECT_KNOWLEDGE, benchmark fixture, skill examples
 4. **Harness policy docs** — CONTRIBUTING, AGENTS, SKILL_INSTALLATION, maintenance playbook
-5. **Doc anti-drift** — `test_routing_contract.py`, cheatsheet parity tests, `test_readme_governance.py` (190+ pytest anti-drift suite in CI)
+5. **Doc anti-drift** — `test_routing_contract.py`, cheatsheet parity tests, `test_readme_governance.py`, `test_thinking_prompts_eval_harness.py` (200+ pytest anti-drift suite in CI)
 
 ### Ongoing maintenance (not blockers)
 

reflective-prompt-library/plans/multi-agent-panel-consensus-2026-06-25.md

@@ -1475,3 +1475,56 @@ User directive: **review all docs** — reopen six-lens panel for cross-surface
 
 _Ongoing maintenance: [GLOSSARY.md](../GLOSSARY.md) Governance Maintenance Playbook — expand ROUTE-002/003 holdout before `route_paraphrase_eval.py` tuning (R8–R12)._
 
+
+
+## Round 69 — Thinking prompt contract review (2026-06-25)
+
+User directive: review prompts, plans, skills, and Socratic/critical-thinking lenses in parallel until all roles agree, then implement.
+
+### DA: Standardize `01-thinking/` prompt contracts?
+
+| Lens | Position |
+| --- | --- |
+| Opus | **Agree** — eval_harness fails without document-level Purpose/Scope/Acceptance/Falsifiability |
+| Codex | **Agree** — map each lens to frozen workflow skills; no tenth skill |
+| Gemini | **Agree** — keep zh-TW templates; add English contract outside code block |
+| Composer | **Agree** — bounded to five thinking prompts, not all 39 prompts |
+| Sakana | **Agree** — aligns with evidence-over-confidence in PROJECT_KNOWLEDGE |
+| GLM | **Agree** — skill-map already routes thinking lenses to brief/review |
+
+**Socratic Q:** Does this duplicate skills?
+**Answer:** No — prompts are composable lenses; skills are workflow contracts. Purpose sections name routing, not replace SKILL.md.
+
+**Consensus:** **Agree** — add Purpose/Scope/Acceptance Criteria/Falsifiability to all five `01-thinking/` prompts; add `test_thinking_prompts_eval_harness.py`.
+
+### DB: Expand to all prompt categories now?
+
+| Lens | Position |
+| --- | --- |
+| All six | **Reject** — recurrence-gated; 01-thinking is highest-leverage for this review |
+
+**Consensus:** **Reject** — defer 00-core/02-engineering Purpose sweep.
+
+### Round 69 verdict table
+
+| ID | Option | Verdict | Action |
+| --- | --- | --- | --- |
+| DA | Thinking prompt contracts | **Agree** | 5 files + pytest anti-drift |
+| DB | All-category Purpose sweep | **Reject** | backlog |
+
+**All roles agree.**
+
+## Implemented Changes (Round 69)
+
+- `01-thinking/*.md`: Purpose, Scope, Acceptance Criteria, Falsifiability + workflow skill mapping
+- `plans/tests/test_thinking_prompts_eval_harness.py`: structural + score floor anti-drift
+- `QUALITY_GATES_SUMMARY.md`: 200+ pytest floor; thinking prompt test mention
+- `PROJECT_KNOWLEDGE.md`: Decision Index Round 69 entry
+
+## Verification (Round 69)
+
+- `make all`: pytest + ROUTE-001/002/003 100%
+
+## Panel status (updated)
+
+**Resealed 2026-06-25** after **Round 69** (options DA–DB). Thinking-prompt contract pass complete; broader prompt-category sweep remains recurrence-gated.

reflective-prompt-library/plans/tests/test_thinking_prompts_eval_harness.py

@@ -0,0 +1,52 @@
+"""Anti-drift: 01-thinking prompts must satisfy eval_harness structural rubric."""
+
+import sys
+from pathlib import Path
+
+import pytest
+
+sys.path.insert(0, str(Path(__file__).parent.parent))
+
+from eval_harness import EvalHarness  # noqa: E402
+
+THINKING_DIR = Path(__file__).parent.parent.parent / "01-thinking"
+REPO_ROOT = str(Path(__file__).parent.parent.parent.parent)
+MIN_SCORE = 80.0
+
+REQUIRED_HEADINGS = (
+    "## Purpose",
+    "## Scope",
+    "## Acceptance Criteria",
+    "## Falsifiability",
+)
+
+THINKING_PROMPTS = tuple(sorted(THINKING_DIR.glob("*.md")))
+
+
+@pytest.fixture(scope="module")
+def harness() -> EvalHarness:
+    return EvalHarness(repo_root=REPO_ROOT)
+
+
+@pytest.mark.parametrize("prompt_path", THINKING_PROMPTS, ids=lambda p: p.name)
+def test_thinking_prompt_has_contract_headings(prompt_path: Path):
+    text = prompt_path.read_text(encoding="utf-8")
+    preamble = text.split("```", 1)[0]
+    for heading in REQUIRED_HEADINGS:
+        assert heading in preamble, f"{prompt_path.name} missing {heading} outside template block"
+
+
+@pytest.mark.parametrize("prompt_path", THINKING_PROMPTS, ids=lambda p: p.name)
+def test_thinking_prompt_meets_eval_harness_floor(prompt_path: Path, harness: EvalHarness):
+    rel = str(prompt_path.relative_to(REPO_ROOT))
+    result = harness.evaluate_file(rel)
+    assert result["score"] >= MIN_SCORE, (
+        f"{prompt_path.name} eval_harness score {result['score']}% < {MIN_SCORE}%: "
+        f"{[(c['id'], c['result']) for c in result['checks']]}"
+    )
+
+
+def test_thinking_prompts_reference_workflow_skills():
+    for prompt_path in THINKING_PROMPTS:
+        text = prompt_path.read_text(encoding="utf-8")
+        assert "reflective-" in text, f"{prompt_path.name} should map to at least one workflow skill"