worclaude 2.1.0 → 2.2.0

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "worclaude",
-  "version": "2.1.0",
+  "version": "2.2.0",
   "description": "CLI tool that scaffolds a comprehensive Claude Code workflow into any project",
   "type": "module",
   "bin": {

package/templates/agents/universal/verify-app.md

@@ -112,6 +112,9 @@ You will feel the urge to skip checks. These are the excuses — recognize them:
 - **Config/Infrastructure**: validate syntax → dry-run where possible → check env vars
 - **Bug fixes**: reproduce original bug → verify fix → run regression tests
 - **Refactoring**: existing test suite must pass unchanged → diff public API surface
+- **Mobile (iOS/Android)**: clean build → install on simulator/emulator → dump accessibility/UI tree, tap by coords, re-dump to verify → check crash logs (logcat / device console)
+- **Database migrations**: run migration up → verify schema matches intent → run migration down (reversibility) → test against existing data, not just empty DB
+- **Data/ML pipeline**: run with sample input → verify output shape/schema/types → test empty input and NaN/null handling → check row counts in vs out for silent data loss
 
 ## Before Issuing PASS
 

package/templates/core/memory-md.md

@@ -59,3 +59,12 @@ Rule or fact.
 - Debugging solutions — the fix is in the code, the context is in the commit message
 - Anything already in CLAUDE.md
 - Ephemeral task details — current conversation context, in-progress work
+
+These exclusions apply even when explicitly asked to save. If asked to save a PR list or activity summary, ask what was *surprising* or *non-obvious* about it — that is the part worth keeping.
+
+## Memory vs Plans vs Tasks
+
+Memory is for information useful across conversations. Don't use it for:
+- **Plans** — use a plan file or plan mode for implementation strategies within a session
+- **Tasks** — use the task system for tracking work steps within a session
+- **Session state** — current conversation context belongs in the conversation, not memory

package/templates/skills/universal/claude-md-maintenance.md

@@ -100,6 +100,28 @@ The /update-claude-md command helps at session end:
 Always review proposed changes before applying. Not every mistake needs a rule.
 Only add rules for recurring problems.
 
+## The @include Directive
+
+When CLAUDE.md grows beyond the 50-line target, use `@include` to split content
+into separate files while keeping it loadable:
+
+```markdown
+# CLAUDE.md
+## Key Files
+@./docs/conventions.md
+@./docs/api-standards.md
+```
+
+- `@./relative` — relative to the file containing the directive
+- `@~/path` — relative to home directory
+- `@/absolute` — absolute path
+- Works in CLAUDE.md, .claude/CLAUDE.md, .claude/rules/*.md, and CLAUDE.local.md
+- Does NOT work inside code blocks (only in leaf text nodes)
+- Non-existent files are silently ignored; circular references are prevented
+
+This is the recommended alternative to cramming everything into CLAUDE.md.
+Each included file still consumes context budget, so use judiciously.
+
 ## Gotchas
 
 - CLAUDE.md is read as system context, not as a document. Write it as instructions,

package/templates/skills/universal/context-management.md

@@ -118,6 +118,29 @@ WORCLAUDE_HOOK_PROFILE=strict claude
 The default is `standard` if the variable is not set. You don't need to do
 anything for normal development — the default just works.
 
+## Token Budgets (Reference)
+
+Approximate values from Claude Code v2.1.88 source (March 2026). These may change between releases.
+
+| Resource | Budget |
+|---|---|
+| Context window (default) | 200,000 tokens |
+| Context window (Opus 4.6 / Sonnet 4.6 with 1M) | 1,000,000 tokens |
+| Max output tokens (Sonnet 4.6) | 32,000 default, 128,000 upper limit |
+| Max output tokens (Opus 4.6) | 64,000 default, 128,000 upper limit |
+| Tool presence overhead | ~500 tokens (added when any tools are enabled) |
+| Post-compact file restore | 5 files max, 50,000 token budget, 5,000 per file |
+| Post-compact skills restore | 25,000 token budget, 5,000 per skill |
+| Compact summary output | 20,000 tokens max |
+| CLAUDE.md total budget | ~40,000 characters across all loaded instruction files |
+| MEMORY.md | 200 lines / 25,000 bytes |
+
+Practical takeaways:
+- After compaction, only 5 files are restored at 5k tokens each. Structure your work
+  so the most important files are recently read.
+- Skills over 5k tokens get truncated after compaction. Keep skills focused.
+- The 70% rule still applies — these numbers help you estimate when you'll hit it.
+
 ## Gotchas
 
 - Compacting doesn't free as much context as you think. If you've read 20 large files,

package/templates/skills/universal/coordinator-mode.md

@@ -21,6 +21,25 @@ Don't use coordinator patterns when:
 - Steps are strictly sequential with no parallelism
 - The overhead of coordination exceeds the work itself
 
+## Phase-Based Workflow
+
+Most coordinated tasks follow four phases:
+
+| Phase | Who | Purpose |
+|-------|-----|---------|
+| Research | Workers (parallel) | Investigate codebase, find files, understand problem |
+| Synthesis | **You** (coordinator) | Read findings, understand the problem, craft implementation specs |
+| Implementation | Workers | Make targeted changes per spec, commit |
+| Verification | Workers | Test changes work, try to break them |
+
+The synthesis phase is the coordinator's most important job. When workers report
+research findings, **you must understand them before directing follow-up work**.
+Read the findings. Identify the approach. Then write a prompt that proves you
+understood by including specific file paths, line numbers, and exactly what to change.
+
+Never write "based on your findings" or "based on the research." These phrases
+delegate understanding to the worker instead of doing it yourself.
+
 ## Worker Prompt Best Practices
 
 Workers cannot see your conversation. Every prompt must be self-contained: