elliot-stack 1.0.11 → 1.0.13

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "elliot-stack",
-  "version": "1.0.11",
+  "version": "1.0.13",
   "description": "Elliot's skill stack for Claude Code — install via npx elliot-stack@latest",
   "bin": {
     "elliot-stack": "bin/install.cjs"

package/skills/estack-better-title/SKILL.md

@@ -11,10 +11,27 @@ The current session ID is: ${CLAUDE_SESSION_ID}
 
 ## Your task
 
-Suggest **3 descriptive titles** for this chat session based on the conversation so far. The goal is to make sessions easy to find later. Good titles:
-- Summarize **what was accomplished** in the chat, not just the topic
-- List key actions/outcomes separated by dashes, commas, or similar (e.g. "Reverse-Engineering /rename, PR #33165 Comment, and Building /better-title Skill")
-- Are detailed enough that someone skimming a session list can tell exactly what happened
+Suggest **3 descriptive titles** for this chat session based on the conversation so far. The goal is to make sessions easy to find later.
+
+### Before drafting, do this quick mental pass
+
+Ask yourself:
+1. What were the **goals** going in?
+2. What **got built / decided / fixed** that the user might reference later?
+3. What was **noise** — mistakes, abandoned approaches, things they had to redirect you on, dead ends? (These do NOT belong in the title.)
+4. What's **temporary** vs. **lasting**? Temporary states should be mentioned briefly, not led with.
+
+Write the title from those answers, NOT from a chronological recap of the chat.
+
+### Good titles
+
+- Summarize **what was accomplished**, not what was attempted
+- **Title the outputs, not the journey** — exclude mistakes, dead ends, and abandoned approaches. They're not searchable and not useful context for future-you.
+- **Use plain language a future-you would search for.** Avoid jargon like "passthrough," "edge cases," "toggle logic," "auto-cd," "refactored," "overhaul." Say what it does in normal words.
+- **Weight by long-term reference value.** If something will stop mattering in weeks or months (temporary fixes, current-state notes, "until X is available" workarounds), mention it briefly but don't lead with it.
+- List key outputs separated by dashes, commas, or similar (e.g. "Reverse-Engineering /rename, PR #33165 Comment, and Building /better-title Skill")
+- Cover 2-4 main outputs. Resist cramming everything in — first-pass attempts tend to over-include.
+- Are detailed enough that someone skimming a session list can tell exactly what they'd find inside
 - Typically 8-20 words — longer is fine if it adds useful detail
 
 ## Format

package/skills/estack-customer-discovery/SKILL.md

@@ -28,16 +28,22 @@ If they describe their situation without picking a step, route them based on whe
 - Have interviews scheduled but need questions → **Step 3: Interview Execution**
 - Have done interviews and need to make sense of them → **Step 4: Analysis & Signals**
 
-### Step 2: Read the right step file
+### Step 2: Read the right step file — MANDATORY
 
-Each step has its own workflow file with detailed instructions. Read the one you need:
+<CRITICAL>
+You MUST read the step file AND its corresponding reference file BEFORE responding to the user. Do NOT skip this. These files contain the actual frameworks and instructions — without reading them, your output will be generic and wrong.
+</CRITICAL>
 
-- **Step 1: Strategy & Targeting** → Read `steps/step-1-strategy.md`
-- **Step 2: Outreach** → Read `steps/step-2-outreach.md`
-- **Step 3: Interview Execution** → Read `steps/step-3-interviews.md`
-- **Step 4: Analysis & Signals** → Read `steps/step-4-analysis.md`
+Each step has its own workflow file with detailed instructions. **Read the step file first, then read the reference file it points to.** Both reads are required — not optional, not "if you have time."
 
-The step file will tell you to also read the corresponding reference document. The reference docs are the source of truth — they contain the principles and frameworks. The step files tell you how to work through them with the user.
+| Step | Step file (read first) | Reference file (read second) |
+|---|---|---|
+| 1. Strategy & Targeting | `steps/step-1-strategy.md` | `references/01-strategy-and-targeting.md` |
+| 2. Outreach | `steps/step-2-outreach.md` | `references/02-outreach.md` |
+| 3. Interview Execution | `steps/step-3-interviews.md` | `references/03-interview-execution.md` |
+| 4. Analysis & Signals | `steps/step-4-analysis.md` | `references/04-analysis-and-signals.md` |
+
+The reference docs are the source of truth — they contain the principles and frameworks. The step files tell you how to work through them with the user.
 
 ### Step 3: Be adaptive
 

package/skills/estack-customer-discovery/steps/step-1-strategy.md

@@ -1,8 +1,10 @@
 # Step 1: Strategy & Targeting
 
-## Before anything else
+## Before anything else — BLOCKING REQUIREMENT
 
-Read `references/01-strategy-and-targeting.md` in the skill directory. That's your source of truth for this step. Internalize it before working with the user.
+<CRITICAL>
+You MUST read `references/01-strategy-and-targeting.md` BEFORE responding to the user. Do NOT skip this. This file contains the actual frameworks and instructions — without reading it, your output will be generic and wrong.
+</CRITICAL>
 
 ## What this step accomplishes
 

package/skills/estack-customer-discovery/steps/step-2-outreach.md

@@ -1,8 +1,10 @@
 # Step 2: Outreach
 
-## Before anything else
+## Before anything else — BLOCKING REQUIREMENT
 
-Read `references/02-outreach.md` in the skill directory. That's your source of truth for this step.
+<CRITICAL>
+You MUST read `references/02-outreach.md` BEFORE responding to the user. Do NOT skip this. This file contains the actual frameworks and instructions — without reading it, your output will be generic and wrong.
+</CRITICAL>
 
 ## What this step accomplishes
 

package/skills/estack-customer-discovery/steps/step-3-interviews.md

@@ -1,8 +1,10 @@
 # Step 3: Interview Execution
 
-## Before anything else
+## Before anything else — BLOCKING REQUIREMENT
 
-Read `references/03-interview-execution.md` in the skill directory. That's your source of truth for this step.
+<CRITICAL>
+You MUST read `references/03-interview-execution.md` BEFORE responding to the user. Do NOT skip this. This file contains the actual frameworks and instructions — without reading it, your output will be generic and wrong.
+</CRITICAL>
 
 ## What this step accomplishes
 

package/skills/estack-customer-discovery/steps/step-4-analysis.md

@@ -1,8 +1,10 @@
 # Step 4: Analysis & Signals
 
-## Before anything else
+## Before anything else — BLOCKING REQUIREMENT
 
-Read `references/04-analysis-and-signals.md` in the skill directory. That's your source of truth for this step.
+<CRITICAL>
+You MUST read `references/04-analysis-and-signals.md` BEFORE responding to the user. Do NOT skip this. This file contains the actual frameworks and instructions — without reading it, your output will be generic and wrong.
+</CRITICAL>
 
 ## What this step accomplishes