chore(compat): ship 8 CC 2.1.143 opt-ins + fix latent // strip; tested_up_to 2.1.132 → 2.1.150 - #56
Merged
Conversation
… bump tested_up_to to 2.1.150 SchemaStore PR #5706 merged 2026-05-23 (sync to CC v2.1.143), unblocking nine settings.json keys + five env vars held in the configurator backlog. This PR ships all of them as commented opt-ins (the PRs #17/#18 pattern) and bumps tested_up_to 2.1.132 → 2.1.150 after the 2.1.147-2.1.150 resurvey found only one new key (allowAllClaudeAiMcps, 2.1.149 — Enterprise managed, outside configurator territory). Latent-bug fix: configure.py was stripping ALL top-level //-prefixed keys when merging settingsPatch files, including the opt-in stubs (// sandbox, // prUrlTemplate) PRs #17 and #18 promised would surface in users' generated .claude/settings.json. New _is_doc_label helper keeps stub keys, drops only label keys (//, //N). Two source patches also corrected (slop-scan, microbit-enforcer) — their top-level documentation keys would otherwise leak under the new filter. Settings-key opt-ins shipped (commented; user uncomments to activate): - skillOverrides → token-efficiency tier-pro patch - worktree.baseRef + worktree.bgIsolation → new multi-agent patch - autoMode.hard_deny + sandbox.failIfUnavailable → safety patch - subagentStatusLine + statusLine.hideVimModeIndicator → ui extraSettings - Hook args/continueOnBlock → doc note in safety + git-workflow patches - CLAUDE_CODE_STOP_HOOK_BLOCK_CAP → safety patch // env block Still held for a future schemastore sync: terminalSequence hook output (2.1.147), Stop/SubagentStop input background_tasks/session_crons (2.1.149). Both verified absent from #5706. Closes #53. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
6 tasks
tigers1997
added a commit
that referenced
this pull request
May 23, 2026
… defaults (#57) Follow-up to PR #56's unblock batch. Two of the eight opt-ins shipped as stubs in that PR are now active defaults, based on a per-key review against the configurator's existing safety/efficiency goals. The other six stay stubbed for genuine reasons (user-workflow choice, infrastructure-dependent, or would break shipped scripts). Promoted: - skillOverrides: "name-only" → active in templates/token-efficiency/settings-patch.tier-pro.json (renders only when token-efficiency.tier=pro). Extends the pro tier's existing "aggressive context efficiency" contract by collapsing skill descriptions — the largest single contributor to per-turn context overhead, the metric /check-context flags. Model still sees skill names. basic tier unaffected. Personas affected: solo-experienced, small-team. Override by editing the patch. - autoMode.hard_deny: ["Running executable files", "Writing to system directories"] → active in templates/safety/settings-patch.json. Pure upside: zero behavior change for standard manual sessions (auto-mode classifier doesn't fire without --auto-mode), meaningful safety backstop for --auto-mode users. Consistent with the configurator's existing safety-first posture (disableBypassPermissionsMode, PreToolUse Bash blocking, scan-secrets, slop-scan). All non-custom personas affected. Tune via the patch. No tested_up_to bump (stays 2.1.150 from PR #56). No new tests required — schema validation + persona snapshots + 50+ shell fixtures all pass. Retrofit impact: existing users on cc-configure --retrofit will gain a top-level autoMode block (all non-custom personas) and, for pro-tier users, a top-level skillOverrides: "name-only". Both are no-ops in common workflows. Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
Merged
3 tasks
tigers1997
added a commit
that referenced
this pull request
May 23, 2026
…ive-default promotions (#58) Bundle release of PRs #56 + #57: - PR #56 — chore(compat): ship 8 schema-validated opt-ins + fix latent // strip; bump tested_up_to 2.1.132 → 2.1.150. Closed tracking issue #53 after a three-resurvey watch on SchemaStore PR #5706 (merged 2026-05-23, sync to CC v2.1.143). - PR #57 — feat(defaults): promote skillOverrides + autoMode.hard_deny to active defaults. Six other opt-ins from #56 stay stubbed for genuine reasons documented in the patch comments. CC_VERSION 2.5.0 → 2.6.0. Claude Code compat 2.1.116–2.1.150. Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
6 tasks
tigers1997
added a commit
that referenced
this pull request
May 24, 2026
…lOverrides (#60) Dogfood from upgrading an adjacent project to the latest configurator version surfaced two real Claude Code settings-validator complaints: 1. `templates/token-efficiency/settings-patch.tier-pro.json` shipped `"skillOverrides": "name-only"` (string form), but the current Claude Code schema requires the per-skill object map. The string form was based on a comment in the patch file that claimed a global "apply to all skills" form existed; verification against the official doc (code.claude.com/docs/en/settings) shows it never did. Also: the setting explicitly doesn't apply to plugin skills, further narrowing its usefulness as a tier-wide default. 2. The `// foo` "commented opt-in stub" pattern propagated literal top-level keys into the user's `.claude/settings.json`. PR #56's `_is_doc_label` filter intentionally kept `// foo` stubs (stripping only bare `//` and `//<digits>`) so users could uncomment them. Empirically, those propagated keys trigger schema-validator complaints in the user's editor — the intent was wrong. A third bug was found during verification: `compute_merged_settings` applies the doc-label filter to file-based patches but bypasses inline `extraSettings` (ui module). So the ui module's bare `"//"` doc-labels were also leaking to user settings.json. Fixes: - `_is_doc_label` now strips ALL `//`-prefixed keys. - New `_strip_doc_labels()` recursively scrubs merged settings — catches nested cases (e.g., `statusLine.// hideVimModeIndicator`) the shallow per-merge filters miss. - `extraSettings` merge path now applies the filter. - `skillOverrides: "name-only"` deleted from tier-pro (replaced with a docstring explaining why no default is shipped). - All `// foo` stubs deleted from 4 source patch files + the ui module's inline `extraSettings`. - Opt-in discovery moved to `templates/core/dot-claude/settings.local.json.example` — the `.example` suffix means Claude Code doesn't parse it directly, so `// foo` stubs there are safe. - New `check_settings_validates()` runtime preflight emits a `[ SETTINGS WARNINGS ]` block when violations are detected. - New static `--check` step asserts no patch file ships `skillOverrides` as a non-object (regression guard — fires on the current bug when first run). New tests under `test/schema-hygiene/`: - All 5 personas render settings.json with zero `//`-prefixed keys - No settings-patch ships `skillOverrides` as a non-object - `check_settings_validates` catches 4 violation classes + clean case - `_strip_doc_labels` recursively strips at all depths Compat: existing users get the cleanup on next `cc-configure --retrofit`. Deep-merge preserves their customizations; the new strip removes leaks. Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
SchemaStore PR #5706 merged 2026-05-23 (sync to CC v2.1.143), unblocking the nine settings.json keys + five env vars tracked in #53. This PR ships all of them as commented opt-ins (the PRs #17/#18 pattern), bumps
tested_up_to2.1.132 → 2.1.150 after a final 2.1.147–2.1.150 resurvey, and fixes a latent strip bug that had been silently dropping opt-in stubs from generated.claude/settings.jsonsince the initial commit.What's in this PR
Settings-key opt-ins (commented; user uncomments to activate, matching the PRs #17
prUrlTemplateand #18sandbox.network.deniedDomainsprecedent):skillOverrides→templates/token-efficiency/settings-patch.tier-pro.json, behind the existingtier=proflagworktree.baseRef+worktree.bgIsolation→ newtemplates/multi-agent/settings-patch.json(multi-agentmodule previously shipped no settings; addedsettingsPatchreference)autoMode.hard_deny+sandbox.failIfUnavailable→templates/safety/settings-patch.json(failIfUnavailable nested inside the existing// sandboxblock)subagentStatusLine+statusLine.hideVimModeIndicator→uimoduleextraSettingsinconfig_schema.pyargs: string[]+continueOnBlock: boolean→ doc note insafetyandgit-workflowsettings-patch.json comment blocks (no behavioral change; forward-looking enablement for user-authored hooks)Env vars:
CLAUDE_CODE_STOP_HOOK_BLOCK_CAPshipped insafety/settings-patch.json's commented// envblock (seeded value"16"). Four others (ANTHROPIC_WORKSPACE_ID,CLAUDE_CODE_OPUS_4_6_FAST_MODE_OVERRIDE,CLAUDE_CODE_PLUGIN_PREFER_HTTPS,CLAUDE_CODE_POWERSHELL_RESPECT_EXECUTION_POLICY) are platform-/workspace-specific and captured in theCLAUDE_CODE_COMPATrationale comment rather than as template defaults.Compat bump:
CLAUDE_CODE_COMPAT.tested_up_to2.1.132 → 2.1.150 with a rewritten rationale comment. 2.1.147–2.1.150 resurvey found one new configurator-adjacent key (allowAllClaudeAiMcps, 2.1.149) — Enterprise managed setting, outside configurator territory.Latent-bug fix
While wiring the new opt-ins I discovered that
configure.py:463was stripping ALL top-level//-prefixed keys when mergingsettingsPatchfiles, including the opt-in stubs (// sandbox,// prUrlTemplate) that PRs #17 and #18 promised users would see in their.claude/settings.json. The filter intended to drop documentation labels ("//", "//2", "//9") but was too broad.New
_is_doc_label(k)helper distinguishes label keys (key is"//"exactly, or//followed by digits only) from stub keys ("// foo"with a name). After the fix:// sandbox,// prUrlTemplate,// worktree,// autoMode,// env,// skillOverrides,// subagentStatusLine,// hideVimModeIndicator,// failIfUnavailableall surface in scaffolded settings.json"//", "//2", "//9"documentation strings are still filtered outSide effect: users on existing scaffolds will see the prUrlTemplate + sandbox.network.deniedDomains stubs in their settings.json on next
cc-configure --retrofit— matching the documented PRs #17/#18 behavior for the first time. No persona snapshot drift; all 50+ shell fixtures still pass.Two source patch files corrected (
templates/safety/settings-patch.slop-scan.json,templates/commands/microbit-enforcer/settings-patch.json) — their top-level"// settings-patch.slop-scan.json"/"// microbit-enforcer.json"keys were maintainer-facing docs that would otherwise leak into user settings under the new filter; rewritten as plain"//"labels.Still held for a future schemastore sync
terminalSequencefield (CC 2.1.147) — NOT in #5706, verified absent from head-branch schema 2026-05-21background_tasks+session_cronsfields (CC 2.1.149) — NOT in #5706Test plan
python3 configure.py --check→ all checks passed; compat now reads2.1.116–2.1.150solo-newer,solo-experienced,small-team,library-author,custom) matchexamples/persona-*/expected-tree.txt--modules core,$modsmoke combos pass (multi-agent,github-actions,lockdown,experiments-memory,mcp,ui).claude/settings.jsonforsmall-teampersona — all 8 new//opt-ins present and well-formedcheckworkflow) passes on this branchCloses #53.
🤖 Generated with Claude Code