@vodailoc/kilo-kit-mcp 1.1.0 → 1.1.1

package/skills/kilo-kit/references/patterns.md

@@ -1,156 +1,156 @@
-# 📐 Patterns Reference
-
-> Reusable patterns for evidence-based AI agent workflows in Kilo-Kit.
-
----
-
-## Pattern 1: Pre-Flight System Scan
-
-**Context:** Before any task execution, verify system readiness.
-
-**Structure:**
-
-```yaml
-pre_flight:
-  steps:
-    - name: Check disk space
-      command: "df -h ."
-      fail_if: "usage > 90%"
-    - name: Check memory
-      command: "free -m"
-      fail_if: "available < 256MB"  # Minimum for running build tools + tests concurrently
-    - name: Check runtime versions
-      command: "node --version && python3 --version"
-      fail_if: "version below minimum"
-    - name: Check running services
-      command: "ps aux | grep -E '(node|python|dotnet)'"
-      purpose: "Identify conflicts"
-```
-
-**When to use:** At the start of every Hard-Gate check.
-
----
-
-## Pattern 2: Codebase Context Gathering
-
-**Context:** Understand the project before making changes.
-
-**Structure:**
-
-```yaml
-codebase_scan:
-  steps:
-    - name: Project structure
-      command: "find . -maxdepth 2 -type f -not -path './.git/*' | head -60"
-    - name: Recent changes
-      command: "git log --oneline -10"
-    - name: Relevant files
-      action: "Search for files related to the task keywords"
-    - name: Existing tests
-      command: "find . -name '*.test.*' -o -name '*.spec.*' | head -20"
-    - name: Dependencies
-      action: "Read package.json, requirements.txt, or equivalent"
-```
-
-**When to use:** During the codebase scan phase of Hard-Gate.
-
----
-
-## Pattern 3: Evidence-Backed Change Proposal
-
-**Context:** Every recommended change must have traceable evidence.
-
-**Structure:**
-
-```markdown
-## Change Proposal
-
-**File:** `<path/to/file>:<line_number>`
-**Current behavior:** <what the code does now>
-**Proposed change:** <what the code should do>
-
-**Evidence:**
-1. Source: `<file:line>` — <what was observed>
-2. Test gap: `<test file>` — <what is missing>
-3. Impact: <scope of affected components>
-
-**Verification plan:**
-- [ ] Unit test added/updated
-- [ ] Existing tests still pass
-- [ ] Manual verification performed
-```
-
-**When to use:** Before any code modification.
-
----
-
-## Pattern 4: Decision Audit Entry
-
-**Context:** Every significant decision must be logged.
-
-**Structure:**
-
-```yaml
-decision:
-  id: "DEC-<timestamp>"
-  action: "<what was decided>"
-  evidence:
-    - "<file:line or command output that supports this>"
-  alternatives_considered:
-    - option: "<alternative approach>"
-      rejected_because: "<reason>"
-  iron_law_compliance:
-    - law_1_evidence: true
-    - law_2_scan_completed: true
-    - law_3_verification_planned: true
-```
-
-**When to use:** At every routing and execution decision point.
-
----
-
-## Pattern 5: Minimal Change Verification
-
-**Context:** Confirm that changes are focused and minimal.
-
-**Structure:**
-
-```yaml
-change_verification:
-  files_modified: ["<list of files>"]
-  lines_changed: <count>
-  scope_check:
-    - question: "Is every changed file directly related to the task?"
-      answer: "yes|no — <explanation>"
-    - question: "Could any change be removed without breaking the fix?"
-      answer: "yes|no — <explanation>"
-    - question: "Are there unintended side effects?"
-      answer: "yes|no — <explanation>"
-```
-
-**When to use:** Before completing any task (Iron Law #4: Minimal Blast Radius).
-
----
-
-## Pattern 6: Failure Recovery
-
-**Context:** When an execution step fails, recover systematically.
-
-**Structure:**
-
-```yaml
-recovery:
-  failed_step: "<which step failed>"
-  error_output: "<actual error message>"
-  diagnosis:
-    - hypothesis: "<what might have caused it>"
-      evidence: "<supporting data>"
-  recovery_action: "<what to do next>"
-  hard_gate_recheck: true  # Always re-run Hard-Gate after failures
-```
-
-**When to use:** Whenever a behavior or quality gate fails.
-
----
-
-*Patterns Reference v1.0.0 — Kilo-Kit*
+# 📐 Patterns Reference
+
+> Reusable patterns for evidence-based AI agent workflows in Kilo-Kit.
+
+---
+
+## Pattern 1: Pre-Flight System Scan
+
+**Context:** Before any task execution, verify system readiness.
+
+**Structure:**
+
+```yaml
+pre_flight:
+  steps:
+    - name: Check disk space
+      command: "df -h ."
+      fail_if: "usage > 90%"
+    - name: Check memory
+      command: "free -m"
+      fail_if: "available < 256MB"  # Minimum for running build tools + tests concurrently
+    - name: Check runtime versions
+      command: "node --version && python3 --version"
+      fail_if: "version below minimum"
+    - name: Check running services
+      command: "ps aux | grep -E '(node|python|dotnet)'"
+      purpose: "Identify conflicts"
+```
+
+**When to use:** At the start of every Hard-Gate check.
+
+---
+
+## Pattern 2: Codebase Context Gathering
+
+**Context:** Understand the project before making changes.
+
+**Structure:**
+
+```yaml
+codebase_scan:
+  steps:
+    - name: Project structure
+      command: "find . -maxdepth 2 -type f -not -path './.git/*' | head -60"
+    - name: Recent changes
+      command: "git log --oneline -10"
+    - name: Relevant files
+      action: "Search for files related to the task keywords"
+    - name: Existing tests
+      command: "find . -name '*.test.*' -o -name '*.spec.*' | head -20"
+    - name: Dependencies
+      action: "Read package.json, requirements.txt, or equivalent"
+```
+
+**When to use:** During the codebase scan phase of Hard-Gate.
+
+---
+
+## Pattern 3: Evidence-Backed Change Proposal
+
+**Context:** Every recommended change must have traceable evidence.
+
+**Structure:**
+
+```markdown
+## Change Proposal
+
+**File:** `<path/to/file>:<line_number>`
+**Current behavior:** <what the code does now>
+**Proposed change:** <what the code should do>
+
+**Evidence:**
+1. Source: `<file:line>` — <what was observed>
+2. Test gap: `<test file>` — <what is missing>
+3. Impact: <scope of affected components>
+
+**Verification plan:**
+- [ ] Unit test added/updated
+- [ ] Existing tests still pass
+- [ ] Manual verification performed
+```
+
+**When to use:** Before any code modification.
+
+---
+
+## Pattern 4: Decision Audit Entry
+
+**Context:** Every significant decision must be logged.
+
+**Structure:**
+
+```yaml
+decision:
+  id: "DEC-<timestamp>"
+  action: "<what was decided>"
+  evidence:
+    - "<file:line or command output that supports this>"
+  alternatives_considered:
+    - option: "<alternative approach>"
+      rejected_because: "<reason>"
+  iron_law_compliance:
+    - law_1_evidence: true
+    - law_2_scan_completed: true
+    - law_3_verification_planned: true
+```
+
+**When to use:** At every routing and execution decision point.
+
+---
+
+## Pattern 5: Minimal Change Verification
+
+**Context:** Confirm that changes are focused and minimal.
+
+**Structure:**
+
+```yaml
+change_verification:
+  files_modified: ["<list of files>"]
+  lines_changed: <count>
+  scope_check:
+    - question: "Is every changed file directly related to the task?"
+      answer: "yes|no — <explanation>"
+    - question: "Could any change be removed without breaking the fix?"
+      answer: "yes|no — <explanation>"
+    - question: "Are there unintended side effects?"
+      answer: "yes|no — <explanation>"
+```
+
+**When to use:** Before completing any task (Iron Law #4: Minimal Blast Radius).
+
+---
+
+## Pattern 6: Failure Recovery
+
+**Context:** When an execution step fails, recover systematically.
+
+**Structure:**
+
+```yaml
+recovery:
+  failed_step: "<which step failed>"
+  error_output: "<actual error message>"
+  diagnosis:
+    - hypothesis: "<what might have caused it>"
+      evidence: "<supporting data>"
+  recovery_action: "<what to do next>"
+  hard_gate_recheck: true  # Always re-run Hard-Gate after failures
+```
+
+**When to use:** Whenever a behavior or quality gate fails.
+
+---
+
+*Patterns Reference v1.0.0 — Kilo-Kit*

package/skills/kilo-kit/references/performance-benchmarks.md

@@ -1,90 +1,90 @@
-# 📊 Performance Benchmarks
-
-> Target benchmarks for Hard-Gate scans, skill execution, and quality gates in Kilo-Kit.
-
----
-
-## Hard-Gate Scan Benchmarks
-
-| Scan Phase | Target Duration | Max Acceptable | Notes |
-|------------|-----------------|----------------|-------|
-| System scan (disk, memory, runtime) | < 5s | 15s | Parallel checks recommended |
-| Codebase scan (structure, files) | < 10s | 30s | Depends on project size |
-| Context validation (intent, scope) | < 3s | 10s | Reasoning step |
-| **Total Hard-Gate** | **< 18s** | **55s** | Sum of all phases |
-
----
-
-## Token Budget Benchmarks
-
-| Operation | Typical Tokens | Max Tokens | Mode |
-|-----------|---------------|------------|------|
-| Hard-Gate scan report | 200–500 | 800 | Economy |
-| Iron Law compliance check | 100–300 | 500 | Economy |
-| Skill routing decision | 150–400 | 600 | Standard |
-| Evidence gathering | 500–1500 | 3000 | Standard |
-| Code modification + verification | 1000–3000 | 8000 | Premium |
-| Root cause analysis | 2000–4500 | 10000 | Premium |
-| Security audit | 1500–3500 | 8000 | Critical |
-
----
-
-## Quality Gate Pass Rates
-
-> Target pass rates for well-configured projects.
-
-| Gate | Target Pass Rate | Warning Threshold | Action if Below |
-|------|------------------|-------------------|-----------------|
-| Pre-Execution (intent, budget) | > 95% | < 90% | Improve intent parsing |
-| Per-Behavior (input/output validation) | > 90% | < 80% | Review behavior configs |
-| Post-Execution (tests, build) | > 85% | < 75% | Review test coverage |
-| Pre-Completion (verification) | > 98% | < 95% | Strengthen verification steps |
-
----
-
-## Codebase Scan Scaling
-
-> Expected scan duration by project size.
-
-| Project Size | Files | Scan Duration (target) | Scan Duration (max) |
-|-------------|-------|----------------------|---------------------|
-| Small | < 100 files | < 3s | 10s |
-| Medium | 100–1000 files | < 10s | 30s |
-| Large | 1000–10000 files | < 30s | 90s |
-| Monorepo | > 10000 files | < 60s | 180s |
-
-**Optimization tips for large codebases:**
-- Use `.gitignore`-aware file listing to skip `node_modules`, `dist`, etc.
-- Limit `find` depth to 2 levels for initial scan (matches `SKILL.md` pre-flight pattern; increase to 3 only for monorepos)
-- Use `git diff` to focus on recently changed files
-- Cache project structure between tasks in the same session
-
----
-
-## Skill Execution Benchmarks
-
-| Skill | Typical Duration | Token Usage | Success Rate Target |
-|-------|------------------|-------------|---------------------|
-| `debugging/systematic` | 2–5 min | 1500–4000 | > 80% |
-| `debugging/root-cause` | 5–15 min | 2000–10000 | > 70% |
-| `debugging/verification` | 1–3 min | 500–2000 | > 95% |
-| `quality/code-review` | 3–8 min | 1000–5000 | > 85% |
-| `quality/testing` | 3–10 min | 1500–6000 | > 80% |
-| `development/security` | 5–12 min | 1500–8000 | > 75% |
-| `development/backend` | 5–15 min | 2000–8000 | > 75% |
-
----
-
-## Iron Law Compliance Metrics
-
-| Metric | Target | Measurement Method |
-|--------|--------|--------------------|
-| Evidence citation rate | 100% of recommendations | Count recommendations with/without file:line citations |
-| Hard-Gate completion rate | 100% of tasks | Count tasks that ran Hard-Gate before execution |
-| Verification before completion | 100% of tasks | Count tasks with post-execution verification |
-| Decision trail completeness | > 95% of decisions logged | Audit decision trail entries vs. actions taken |
-| Minimal change adherence | > 90% of tasks | Review changed files vs. task scope |
-
----
-
-*Performance Benchmarks v1.0.0 — Kilo-Kit*
+# 📊 Performance Benchmarks
+
+> Target benchmarks for Hard-Gate scans, skill execution, and quality gates in Kilo-Kit.
+
+---
+
+## Hard-Gate Scan Benchmarks
+
+| Scan Phase | Target Duration | Max Acceptable | Notes |
+|------------|-----------------|----------------|-------|
+| System scan (disk, memory, runtime) | < 5s | 15s | Parallel checks recommended |
+| Codebase scan (structure, files) | < 10s | 30s | Depends on project size |
+| Context validation (intent, scope) | < 3s | 10s | Reasoning step |
+| **Total Hard-Gate** | **< 18s** | **55s** | Sum of all phases |
+
+---
+
+## Token Budget Benchmarks
+
+| Operation | Typical Tokens | Max Tokens | Mode |
+|-----------|---------------|------------|------|
+| Hard-Gate scan report | 200–500 | 800 | Economy |
+| Iron Law compliance check | 100–300 | 500 | Economy |
+| Skill routing decision | 150–400 | 600 | Standard |
+| Evidence gathering | 500–1500 | 3000 | Standard |
+| Code modification + verification | 1000–3000 | 8000 | Premium |
+| Root cause analysis | 2000–4500 | 10000 | Premium |
+| Security audit | 1500–3500 | 8000 | Critical |
+
+---
+
+## Quality Gate Pass Rates
+
+> Target pass rates for well-configured projects.
+
+| Gate | Target Pass Rate | Warning Threshold | Action if Below |
+|------|------------------|-------------------|-----------------|
+| Pre-Execution (intent, budget) | > 95% | < 90% | Improve intent parsing |
+| Per-Behavior (input/output validation) | > 90% | < 80% | Review behavior configs |
+| Post-Execution (tests, build) | > 85% | < 75% | Review test coverage |
+| Pre-Completion (verification) | > 98% | < 95% | Strengthen verification steps |
+
+---
+
+## Codebase Scan Scaling
+
+> Expected scan duration by project size.
+
+| Project Size | Files | Scan Duration (target) | Scan Duration (max) |
+|-------------|-------|----------------------|---------------------|
+| Small | < 100 files | < 3s | 10s |
+| Medium | 100–1000 files | < 10s | 30s |
+| Large | 1000–10000 files | < 30s | 90s |
+| Monorepo | > 10000 files | < 60s | 180s |
+
+**Optimization tips for large codebases:**
+- Use `.gitignore`-aware file listing to skip `node_modules`, `dist`, etc.
+- Limit `find` depth to 2 levels for initial scan (matches `SKILL.md` pre-flight pattern; increase to 3 only for monorepos)
+- Use `git diff` to focus on recently changed files
+- Cache project structure between tasks in the same session
+
+---
+
+## Skill Execution Benchmarks
+
+| Skill | Typical Duration | Token Usage | Success Rate Target |
+|-------|------------------|-------------|---------------------|
+| `debugging/systematic` | 2–5 min | 1500–4000 | > 80% |
+| `debugging/root-cause` | 5–15 min | 2000–10000 | > 70% |
+| `debugging/verification` | 1–3 min | 500–2000 | > 95% |
+| `quality/code-review` | 3–8 min | 1000–5000 | > 85% |
+| `quality/testing` | 3–10 min | 1500–6000 | > 80% |
+| `development/security` | 5–12 min | 1500–8000 | > 75% |
+| `development/backend` | 5–15 min | 2000–8000 | > 75% |
+
+---
+
+## Iron Law Compliance Metrics
+
+| Metric | Target | Measurement Method |
+|--------|--------|--------------------|
+| Evidence citation rate | 100% of recommendations | Count recommendations with/without file:line citations |
+| Hard-Gate completion rate | 100% of tasks | Count tasks that ran Hard-Gate before execution |
+| Verification before completion | 100% of tasks | Count tasks with post-execution verification |
+| Decision trail completeness | > 95% of decisions logged | Audit decision trail entries vs. actions taken |
+| Minimal change adherence | > 90% of tasks | Review changed files vs. task scope |
+
+---
+
+*Performance Benchmarks v1.0.0 — Kilo-Kit*