superkit-mcp-server 1.1.4 → 1.2.0

package/skills/workflows/skill-review.md

@@ -1,127 +1,127 @@
----
-description: Weekly review of potential new agent skills discovered from usage patterns
----
-
-# /skill-review - Process Skill Suggestions
-
-Systematically review and promote suggested skills to formal capabilities.
-
-> **Why review skills?** The system automatically suggests new skills based on user requests (e.g., "run tests", "check logs"). Reviewing these suggestions allows the agent to evolve its capabilities organically.
-
-## When To Use
-
-- Weekly (Mondays) as part of system health checks
-- When `skill_suggestions.csv` grows large (>50 lines)
-- Before creating a new skill manually
-
----
-
-## Workflow
-
-### Step 0: Search Before Solving
-
-```bash
-// turbo
-Call MCP `call_tool_logger_manager` { action: "logWorkflow", name: "/skill-review", outcome: "success" }
-Call MCP `call_tool_compound_manager` { action: "search", terms: [ "skill patterns" "agent capabilities"] }
-Call MCP `call_tool_logger_manager` { action: "logSkill", name: "compound-docs", outcome: "workflow" }
-```
-
----
-
-### Step 1: Analyze Suggestions
-
-1. **Check suggestion volume:**
-   ```bash
-   wc -l .agent/logs/skill_suggestions.csv
-   ```
-
-2. **View unique suggestions:**
-   ```bash
-   cut -d',' -f1 .agent/logs/skill_suggestions.csv | sort | uniq -c | sort -nr | head -20
-   ```
-
-3. **Identify candidates:**
-   Look for high-frequency requests that aren't yet skills.
-   - *Example:* "run_tests" (15 times) → Candidate for `testing` skill (or finding why it's not used)
-   - *Example:* "deploy_app" (8 times) → Candidate for `deployment` skill
-
----
-
-### Step 2: Validate Candidates
-
-For each top candidate:
-
-1. **Check if it already exists:**
-   ```bash
-   grep -r "{keyword}" skills/
-   ```
-   *If it exists but isn't being used: Investigate discoverability.*
-
-2. **Check for existing solutions:**
-   ```bash
-   Call MCP `call_tool_compound_manager` { action: "search", terms: [ "{candidate name}"] }
-   ```
-
----
-
-### Step 3: Promote or Archive
-
-#### Option A: Promote to New Skill
-If a clear gap exists:
-
-1. Run `/create-agent-skill`
-2. Name it clearly (e.g., `deployment`, `database-migration`)
-3. Implement core patterns
-
-#### Option B: Alias to Existing Skill
-If it's a synonym for an existing skill:
-
-1. Add keywords to the existing `SKILL.md` description
-2. Update `.agent/workflows/README.md` to reference it
-
-#### Option C: Archive (Low Value)
-If it's noise or one-off:
-- Ignore it.
-
----
-
-### Step 4: Maintenance
-
-**Rotate the log file:**
-
-> [!IMPORTANT]
-> Clear the suggestions log after review to reset the signal-to-noise ratio.
-
-```bash
-# Append to archive
-cat .agent/logs/skill_suggestions.csv >> .agent/logs/skill_suggestions_archive.csv
-
-# Clear active log
-printf "suggestion,context,count\n" > .agent/logs/skill_suggestions.csv
-```
-
----
-
-### Phase 5: Completion & Handoff
-
-#### Step 1: Establish Terminal UI State
-
-```javascript
-await task_boundary({
-  TaskName: "[COMPLETED] Skill Review",
-  TaskStatus: "Suggestions processed. Offering next steps.",
-  Mode: "VERIFICATION",
-  TaskSummary: "Reviewed top skill suggestions. Promoted X, Archived Y. Cleared suggestions log."
-});
-```
-
-#### Step 2: Mandatory Handoff
-
-```bash
-✓ Skill review complete
-
-Next steps:
-1. /create-agent-skill - If you identified a new skill to build
-2. /housekeeping - If finished
-```
+---
+description: Weekly review of potential new agent skills discovered from usage patterns
+---
+
+# /skill-review - Process Skill Suggestions
+
+Systematically review and promote suggested skills to formal capabilities.
+
+> **Why review skills?** The system automatically suggests new skills based on user requests (e.g., "run tests", "check logs"). Reviewing these suggestions allows the agent to evolve its capabilities organically.
+
+## When To Use
+
+- Weekly (Mondays) as part of system health checks
+- When `skill_suggestions.csv` grows large (>50 lines)
+- Before creating a new skill manually
+
+---
+
+## Workflow
+
+### Step 0: Search Before Solving
+
+```bash
+// turbo
+Call MCP `call_tool_logger_manager` { action: "logWorkflow", name: "/skill-review", outcome: "success" }
+Call MCP `call_tool_compound_manager` { action: "search", terms: [ "skill patterns" "agent capabilities"] }
+Call MCP `call_tool_logger_manager` { action: "logSkill", name: "compound-docs", outcome: "workflow" }
+```
+
+---
+
+### Step 1: Analyze Suggestions
+
+1. **Check suggestion volume:**
+   ```bash
+   wc -l .agent/logs/skill_suggestions.csv
+   ```
+
+2. **View unique suggestions:**
+   ```bash
+   cut -d',' -f1 .agent/logs/skill_suggestions.csv | sort | uniq -c | sort -nr | head -20
+   ```
+
+3. **Identify candidates:**
+   Look for high-frequency requests that aren't yet skills.
+   - *Example:* "run_tests" (15 times) → Candidate for `testing` skill (or finding why it's not used)
+   - *Example:* "deploy_app" (8 times) → Candidate for `deployment` skill
+
+---
+
+### Step 2: Validate Candidates
+
+For each top candidate:
+
+1. **Check if it already exists:**
+   ```bash
+   grep -r "{keyword}" skills/
+   ```
+   *If it exists but isn't being used: Investigate discoverability.*
+
+2. **Check for existing solutions:**
+   ```bash
+   Call MCP `call_tool_compound_manager` { action: "search", terms: [ "{candidate name}"] }
+   ```
+
+---
+
+### Step 3: Promote or Archive
+
+#### Option A: Promote to New Skill
+If a clear gap exists:
+
+1. Run `/create-agent-skill`
+2. Name it clearly (e.g., `deployment`, `database-migration`)
+3. Implement core patterns
+
+#### Option B: Alias to Existing Skill
+If it's a synonym for an existing skill:
+
+1. Add keywords to the existing `SKILL.md` description
+2. Update `.agent/workflows/README.md` to reference it
+
+#### Option C: Archive (Low Value)
+If it's noise or one-off:
+- Ignore it.
+
+---
+
+### Step 4: Maintenance
+
+**Rotate the log file:**
+
+> [!IMPORTANT]
+> Clear the suggestions log after review to reset the signal-to-noise ratio.
+
+```bash
+# Append to archive
+cat .agent/logs/skill_suggestions.csv >> .agent/logs/skill_suggestions_archive.csv
+
+# Clear active log
+printf "suggestion,context,count\n" > .agent/logs/skill_suggestions.csv
+```
+
+---
+
+### Phase 5: Completion & Handoff
+
+#### Step 1: Establish Terminal UI State
+
+```javascript
+await task_boundary({
+  TaskName: "[COMPLETED] Skill Review",
+  TaskStatus: "Suggestions processed. Offering next steps.",
+  Mode: "VERIFICATION",
+  TaskSummary: "Reviewed top skill suggestions. Promoted X, Archived Y. Cleared suggestions log."
+});
+```
+
+#### Step 2: Mandatory Handoff
+
+```bash
+✓ Skill review complete
+
+Next steps:
+1. /create-agent-skill - If you identified a new skill to build
+2. /housekeeping - If finished
+```