C-129 evidence: take and drop value domain, negative asymmetry pinned

[O6lvl4]

Greenfield stage-63 catch: list.take's VALUE had no observer in the claimed corpus, and a min-select operand inversion (take computed MAX — returned the whole list) sat latent for eighteen stages until heap_result_if_bind's chain opened.

• The fixture pins the take/drop value domain from both spellings: in-range, count 0, past-len, and the NEGATIVE reinterpretation asymmetry (take(-1) = whole list, drop(-1) = empty — v0's n as usize verbatim, deliberately not the C-054 clamp).
• Verified on native 0.58.0 (output above the fold in the fixture comment); greenfield claims the row at the next pin bump (2-PR flow). Greenfield-side closure already landed: fix + fuzz arms (take/drop/insert with negative counts) + mutant 040 pinning the exact inversion.
• contracts.toml evidence row under C-129; conformance.md/README regenerated; check-contracts.sh and check-als-style.sh green (601/601 linked).

🤖 Generated with Claude Code