C-Level Advisory DecisionsMemoryTracking

Decision Logger

Two-layer memory architecture for board meeting decisions.

$skillfish add ~skills/decision-logger

Capabilities Real Examples Get Started Models & Context

What This Skill Can Do

Concrete capabilities you get when you install this skill.

Capture and structure board and executive decisions with full context and rationale

Build searchable decision archives with outcome tracking and retrospectives

Create decision templates for recurring decision types (hiring, build/buy, pricing)

Design pre-mortem and post-decision review processes

Track decision outcomes and feed learnings back into future decisions

Maintain institutional memory across leadership team changes

Real Conversations

See how the Decision Logger skill responds to actual requests. These are representative examples of what your AI agent will produce.

You

We just had our Q4 board meeting — 3 hours, 6 major decisions, 4 action items. I took rough notes but by tomorrow morning half the context will be lost. Two board members are already sending conflicting follow-up emails about what was agreed. What's my process?

Agent

Conflicting follow-up emails within 24 hours of a board meeting is a documentation gap, not a disagreement. Here's the fix.

Immediate Recovery (next 2 hours):

For each of your 6 decisions, capture:

Decision: [What was decided, specifically]
Context: [What information drove the decision]
Who decided: [Board vote or consensus?]
Alternatives rejected: [What was considered and why it lost]
Owner: [Who executes]
Deadline: [By when]
Review trigger: [What would cause us to revisit this?]

Send a "Decision Memo" within 24 hours: Not meeting minutes. A decision memo. One page maximum, structured as:

"In our Q4 board meeting on [date], we made the following decisions:"
6 numbered decisions in the format above
"Please reply by [48 hours] with any corrections. Silence implies agreement."

This creates a durable record AND forces resolution of any conflicting interpretations before they calcify.

Two-Layer Memory System for Ongoing Board Decisions:

Session layer: Decision memo per meeting
Persistent layer: Running decision register in Notion/Confluence — every board decision, searchable, with status (open/implemented/reversed)

For next board meeting:

Open with "Decisions from last meeting and their status" — 5 minutes, accountability built in
Assign a decision scribe (not the CEO) whose job is capturing decisions real-time

Do you want a template for the decision memo and the running register?

You

We made a bad hire at VP level 8 months ago that cost us ~$180K and 6 months of momentum. Looking back, we made a similar mistake 2 years ago. We clearly have a pattern but no system to learn from it. How do I build decision memory that actually improves future decisions?

Agent

Two bad VP hires with the same failure pattern suggests a systematic bias in your decision process — and without a logging system, you'll make it a third time.

What Decision Memory Needs to Capture:

For high-stakes decisions (VP+ hires, strategic pivots, major investments):

Decision: VP of [Role] — hire [Name]
Date: [Date]
Context: What problem were we solving? What was the urgency level?
Information we had: References, interview panel scores, concerns raised
Information we lacked: What did we NOT know that would have changed the decision?
Assumptions made: "We assumed they could build from scratch"
Dissenting voices: Did anyone object? What did they say?
Outcome: [8 months later — what happened]
Root cause: Which assumption was wrong?
What we'd do differently: [Specific]

Pattern Analysis — Your Two VP Hires: Look at both failure logs and ask:

Were the roles similar (player-coach vs. pure executive)?
Were the interview processes the same?
Did both hires have strong pedigree but unproven founder-stage experience?
Were there dissenting voices both times that were overridden?

The Structural Fix for VP+ Hires:

Pre-mortem required: "Imagine it's 12 months from now and this hire failed. What happened?"
Contrarian interview: assign one interviewer whose job is to find reasons NOT to hire
Reference check specifically on the failure mode you're worried about (not generic references)

What does the pattern look like between the two failures?

Get Started

How to install and use this skill in your preferred environment.

Skills are designed for AI coding agents (Claude Code, Cursor, Windsurf) and IDE-based workflows where the agent can read files, run scripts, and act on your codebase. Web-based AI can use the knowledge and frameworks, but won't have tool access.

Models & Context

Which AI models and context windows work best with this skill.

Recommended Models

Best

Claude Opus 4 Claude Sonnet 4 GPT-4.1 Gemini 2.5 Pro Grok 3 Kimi K2

Good

Claude Haiku 4.5 GPT-4.1 mini Gemini 2.5 Flash Grok 3 mini

Larger models produce more detailed, production-ready outputs.

Context Window

This skill's SKILL.md is typically 3–10 KB — fits in any modern context window.

8K Skill only

32K+ Skill + conversation

100K+ Skill + references + codebase

All current frontier models (Claude, GPT, Gemini) support 100K+ context. Use the full window for complex multi-service work.

Pro tips for best results

1

Be specific

Include numbers — users, budget, RPS — so the skill can size the architecture.

2

Share constraints

Compliance needs, team size, and existing stack all improve the output.

3

Iterate

Start with a high-level design, then ask follow-ups for IaC, cost analysis, or security review.

4

Combine skills

Pair with companion skills below for end-to-end coverage.

Ready to try Decision Logger?

Install the skill and start getting expert-level guidance in your workflow — any agent, any IDE.