feat(skills): add one-three-one-rule communication skill

skills/communication/one-three-one-rule/SKILL.md

@@ -0,0 +1,94 @@
+---
+name: one-three-one-rule
+description: Enforces the 1-3-1 communication framework for clear, concise, and structured agent responses.
+version: 1.0.0
+author: Willard Moore
+license: MIT
+category: communication
+metadata:
+  hermes:
+    tags: [Communication, Decision-Making, Structure]
+---
+
+# 1-3-1 Communication Rule
+
+This skill ensures all responses adhere to a strict 1-3-1 format: Problem, Options, Recommendation, followed by Definition of Done (DoD) and Implementation Plan tailored to the chosen option.
+
+## When to Use
+
+- When the user explicitly asks for a "1-3-1" response.
+- When a task requires a structured decision with trade-offs and a clear recommendation.
+- When you need to keep communication focused, concise, and action-oriented.
+
+## Procedure (Rules)
+
+1.  **Problem:**
+    *   Must be a single, concise sentence.
+    *   Must state the desired outcome or the core issue.
+    *   Must NOT contain the word "and".
+    *   Must NOT mention implementation details (e.g., specific tools, data sources).
+    *   Example: `Linear issue tickets must be accessible for immediate work.`
+
+2.  **Options:**
+    *   Present exactly three distinct, viable options for solving the problem.
+    *   Each option (A, B, C) should have a brief description, pros, and cons.
+
+3.  **Recommendation:**
+    *   Clearly state the single best option based on current context, user priorities, or efficiency.
+    *   This is the agent's primary suggestion.
+
+4.  **Definition of Done (DoD):**
+    *   This section details the specific success criteria for the **recommended** option.
+    *   It outlines what constitutes a "completed" task for that particular path.
+    *   *Note:* If the user selects a different option, the DoD must be updated to reflect the success criteria for *that* newly recommended option.
+
+5.  **Implementation Plan:**
+    *   This section details the concrete steps required to achieve the **recommended** option.
+    *   It outlines the tooling, commands, and workflow.
+    *   *Note:* Similar to the DoD, the Implementation Plan must be updated if a different option becomes the recommendation.
+
+## Verification
+
+- The response includes exactly one **Problem** sentence with no "and".
+- The response includes exactly three **Options** (A, B, C) with pros and cons.
+- The response includes a single **Recommendation** that clearly picks one option.
+- The response includes a **Definition of Done** and **Implementation Plan** aligned with the recommended option.
+- If the user picks a different option, the updated response adjusts Recommendation, DoD, and Implementation Plan to that option.
+
+## Example Interaction
+
+User: I need to integrate with Linear.
+
+Agent:
+1) Problem: Linear issue tickets must be accessible for immediate work.
+
+2) Options:
+   - Option A: Full GraphQL bootstrap (projects + all issues + all milestones) in one pass.
+     Pros: Complete initial state; one-shot setup.
+     Cons: More data fetched upfront; longer bootstrap.
+   - Option B: GraphQL bootstrap (projects + issues) + periodic milestone refresh.
+     Pros: Fast initial load; data stays fresh with regular refresh.
+     Cons: Requires a scheduler; more moving parts.
+   - Option C: GraphQL bootstrap (projects + issues) + on-demand milestones (no ongoing refresh).
+     Pros: Fastest path to start work; simplest architecture; milestones retrieved only when needed.
+     Cons: Milestones not pre-fetched; potential on-demand latency if needed soon after bootstrap.
+
+3) Recommendation: Option C. This is the simplest, fastest path to get real work started and minimizes upfront effort.
+
+4) Definition of Done (DoD) for Option C:
+   - GraphQL bootstrap retrieves projects and issues, stored in persistent memory.
+   - Memory persists to `~/.hermes/linear_memory.json`.
+   - `graphql.py` includes `get-issues --project-id <id>` for on-demand fetching.
+   - Live proof shows sample project/issue data and bootstrap counts.
+   - Clear path to start work is established.
+
+5) Implementation Plan for Option C:
+   - Step 1: Consolidate into `graphql.py`. Keep `mcp.py` dormant.
+   - Step 2: Configure `~/.hermes/linear_memory.json` for projects and issues.
+   - Step 3: Implement `graphql.py bootstrap`.
+   - Step 4: Implement `graphql.py get-issues`.
+   - Step 5: Provide live proof.
+   - Step 6: Save approach.
+
+### Note on Dynamic Options
+If the user selects Option A or B, the Recommendation, DoD, and Implementation Plan will be adjusted accordingly, based on the chosen option's specifics.