Practice rows: --max-findings, marketplace autoUpdate and trust, the workflow model override, the effortLevel origin - #1147

Scope

• What this PR changes, in one or two sentences: four practice-skill row edits. Three come from the 2026-10-15 guide screen (items 15, 35 and 60); the fourth records the origin of the project's effortLevel "xhigh" (the command center's item Verify rendered dashboards and fix memory panel usability #42).
• Base commit: f165498aa
• Lane: lane:foundation
• Owned paths touched:
  • .claude/skills/claude-native-practice/reference/slots.json and three rendered layer pages;
  • the slot manifest;
  • tests/test_native_practice_ledger_rows_2.py (new);
  • manifests/evidence.json.

Two more items from the same screen land in other open PRs that already edit those rows:

• item 26, the syncClaudeAiSkills wording, goes in the skills PR;
• item 29, the autoMemoryDirectory trust rule, goes in the memory-maintenance PR.

SOTA sources

• anthropics/claude-code CHANGELOG at 94506ff0d3bd9185f1752da36e968775021fb48f:
  • 2.1.288: "Added --max-findings <n>|all to /code-review … the choice is reused until you pass --max-findings default";
  • 2.1.296: "Added CLAUDE_CODE_WORKFLOW_SUBAGENT_MODEL to run every workflow agent on one model while other subagents keep theirs".
• settings-reference, extraKnownMarketplaces (fetched 2026-10-15, client 2.1.300):
  • entries in a repository's settings are honoured "only after you accept the workspace trust dialog … including a -p run there, it ignores them";
  • "When omitted, claude-plugins-official and most other official Anthropic marketplaces default to true, and third-party marketplaces default to false."
• Sonnet 5.5 and Opus 5.5 dispatch: fan-out units on Sonnet 5.5, judgment on Opus 5.5, effort and default child model saved in settings (2.1.284 probes) #479: squash ba31dcd02102b968c696424b9f3966bf0b9334a5, with record docs/decisions/2026-09-29-sonnet-5-5-dispatch.md.

Evidence-class table

Local commands run

$ python3 -m unittest tests.test_native_practice_ledger_rows_2      # before the change
FAILED (failures=4)
$ python3 -m unittest tests.test_native_practice_*  (9 modules)
Ran 105 tests; OK (skipped=3)
$ python3 scripts/validate.py
rc 0

Notes for the reads

Every open practice-row PR (#1142, #1146, this one, and the memory and skills PRs) rewrites the slot manifest's sha256_when_labeled. Each one after the first to land needs a manifest rebase, recomputed from the merged slots.json.

Checklist

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

🤖 Generated with Claude Code