Dispatch record: final status of the unresolved items; the max default carries its caveat in the portable CLAUDE.md - #504

Scope

• What this PR changes, in one or two sentences: refreshes the "Unresolved" section of docs/decisions/2026-09-29-sonnet-5-5-dispatch.md with the final status of each open item after the 2026-09-29 merges, and corrects one parenthetical (open PR Refuse env-style shell launchers; keep unix and local URL schemes checked (#206 review lows) [hold: after the wave] #216 does not touch the GPT-6 lane effort default); and adds one 12-word clause to the portable examples/claude-native/CLAUDE.md so the max default carries the caveat that AGENTS.md, the README and the records already state ("rests on the user's requirement, not on a measured gain"), found by another session's read-only check of main. Docs and one instruction sentence.
• Base commit: 5cfa2400 (origin/main at the last fetch)
• Lane: lane:foundation. The only shared hot file is manifests/evidence.json, by registering the record's new hash in the last commit.
• Owned paths touched: docs/decisions/2026-09-29-sonnet-5-5-dispatch.md, examples/claude-native/CLAUDE.md; manifests/evidence.json (hash registration).
• Not touched: tools/sota-convergence/codex_lane.py, the recipes, catalogs/us-equities/models.json and every other file the items name; each has an owner (below).

SOTA sources

• The facts are read from this repository at the base commit: tools/sota-convergence/codex_lane.py:139 (DEFAULT_EFFORT = "high"), recipes/sota-convergence-practice.md and tools/sota-convergence/README.md (--effort high), scripts/verdict_review_gate.py (TRUST_PATHS), tools/sota-convergence/lane-provenance.json, catalogs/us-equities/models.json (limitations[0]), docs/foundation-stack.md, docs/token-efficiency-stack.json and blueprints/us-equities/routing/README.md (claude/claude-opus-5), and PR Refuse env-style shell launchers; keep unix and local URL schemes checked (#206 review lows) [hold: after the wave] #216's file list; the "49 of 76" count comes from running python3 -m unittest tests.test_codex_lane in a scratch tree with DEFAULT_EFFORT = "max" (removed afterwards).
• The ownership statements come from the sessions that own each area, quoted in their messages of 2026-09-29: native-agent-stack-2d (Gate A, Amendment 4) and ecosystem-roadmap-2026 (the trading lane and the roadmap). No upstream source is installed or followed.

Evidence-class table

Local commands run

$ python3 scripts/validate.py        exit 0  {"components": 69, "hashed_files": 8083, "profiles": 4, "receipts": 171, "status": "passed"}
$ python3 scripts/component_matrix.py --check ; python3 scripts/new_host_grand_list.py --check    exit 0 and exit 0
$ python3 -m unittest tests.test_adoption_docs_consistency tests.test_install_claude_profile     exit 0  Ran 100 tests, OK (skipped=4)
$ node examples/claude-native/workflows/test-envelope.mjs ; node examples/claude-native/workflows/test-contract-mutations.mjs     SUMMARY passed=254 failed=0 ; passed=74 failed=0

Decision record

docs/decisions/2026-09-29-sonnet-5-5-dispatch.md (this PR edits its "Unresolved" section). No review round is claimed: the change records statuses and owners, and adds a caveat to a rule without changing it. The host ~/.claude/CLAUDE.md takes the same sentence after this merges (backup, read-back), untracked.

Host evidence

Not applicable: no files under evidence/hosts/ change.

Checklist

• No new or changed GitHub Actions.
• No new or changed workflows.
• No secrets are printed, logged or committed; no new required secret was added.
• No new paid hosting, subscription or billing surface was introduced.
• Peer-owned untracked files and worktrees were preserved.

🤖 Generated with Claude Code