fix(scaffold): metrics-clean RUNLOG seed — retire both recipes' inline resets

[SollanSystems]

Closes #40.

Root fix, metrics-clean by construction: templates/RUNLOG.md.tmpl no longer seeds a parseable placeholder iteration. The seeded RUNLOG now carries an ## Entry format (reference — not an iteration, not scored) section that keeps the full instructional value (iteration format + complete outcome vocabulary for humans) while emitting zero ## Iteration blocks — metrics._runlog_blocks only scans ^## Iteration headings, so a fresh scaffold scores clean with no fabricated state.

• The issue's acceptance snippet (emit.open_contract(ws) with no args) now prints []; iterations_claiming_success == 0.
• No seed_runlog=False kwarg was added — clean-by-construction makes it dead weight (open_contract(target) signature unchanged; YAGNI, per the option the issue offers).
• Both recipes drop their inline RUNLOG resets (examples/langgraph-emit/graph_example.py, examples/temporal-certify/workflow_example.py) and their explanatory comments; recipe docs/README prose updated where it referenced the reset.
• The v0.8.0 plan's Global-Constraints carve-out paragraph is annotated [Retired …] (historical text preserved).
• scripts/test_template_roundtrip.py drops the 14 now-unused RUNLOG fill keys — its strictest guard (test_every_template_fills_with_no_placeholder_remaining) can only catch more unfilled placeholders after key removal, never fewer.

New regression test in scripts/test_emit.py (test_open_contract_is_metrics_clean): asserts provenance.unrecognized_outcomes == [] and iterations_claiming_success == 0 on a fresh scaffold.

Verification (independent adversarial pass):

• Negative control: the new test overlaid on an unfixed main archive fails with AssertionError: assert ['replace'] == [].
• Full suite with engines installed (langgraph + temporalio): 399 passed / 11 skipped / 0 failed on the branch vs 398/11 on clean main in the identical env — delta is exactly the new test, zero regressions. The recipe ST3 tests assert unrecognized_outcomes == [] on the un-reset scaffold and pass.
• Fresh-scaffold doctor round-trip still clean (test_open_contract_is_doctor_clean); self_eval 13/13; validate_frontmatter 9/0.

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.