skills.rest
tmalcolm-0607
tmalcolm-0607COMMUNITY
Explore all skills by this author →
WebsiteGitHub

council-retro

Capture quantitative learning signals from AI council channel retrospectives.

Updated May 7, 2026
Search Tags:#retrospective#governance#council#ai-agents#correlation#blameless#learning-signals
One-click install
npx skills add https://github.com/tmalcolm-0607/mad-council-claw --skill council-retro
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: council-retro
Source: https://github.com/tmalcolm-0607/mad-council-claw/tree/main/.claude/skills/council-retro
Command: npx skills add https://github.com/tmalcolm-0607/mad-council-claw --skill council-retro

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

AI council channels generate valuable collaboration data during their lifecycle, but without structured retrospectives, lessons about accuracy, process alignment, and team dynamics are lost when channels wind down. This Skill solves that by enforcing a blameless, no-PII reflection ritual that converts subjective experiences into quantitative learning signals correlated with objective outcomes.

Core Features & Use Cases

  • Five-Axis Scoring: Captures 1-5 ratings on accuracy, completeness, TSG alignment, developer experience, and confidence to establish baseline calibration data.
  • Blameless Culture Enforcement: Prompts and storage rules prevent individual attribution and PII exposure, creating safety for honest low-score reporting.
  • Outcome Correlation Engine: Tags each retro with a run ID so future council verdicts and PR quality metrics can be matched against self-assessments to identify miscalibration patterns.

Quick Start

Use the council-retro skill to capture a learning signal for the channel named es-training.

Frequently Asked Questions about council-retro

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a blameless retrospective for multi-agent AI collaboration channels?▼

Run a blameless retrospective by using structured prompts to capture honest learning signals at channel wind-down. This enforces no-PII storage rules and prevents individual attribution, creating safety for accurate low-score reporting in multi-agent contexts.

What is five-axis scoring in AI council post-mortems?▼

Five-axis scoring in AI council post-mortems captures 1-5 ratings on accuracy, completeness, TSG alignment, developer experience, and confidence. This mechanism establishes baseline calibration data to convert subjective collaboration experiences into quantitative metrics.

How do I correlate AI agent self-assessments with objective outcome tracking?▼

Correlate AI agent self-assessments with objective outcomes by tagging each retrospective with a run ID. This allows future council verdicts and PR quality metrics to be matched against self-assessments to identify miscalibration patterns over time.

Does this retrospective process require session-local storage and cross-boundary consent?▼

Yes, the retrospective process requires session-local storage with atomic file writes and cross-boundary consent gates. It also includes prompt-injection scanning and optional ALAS hub submission to satisfy multi-agent governance and security requirements.

What is the best way to capture learning signals when AI agent channels wind down?▼

The best way to capture learning signals during channel wind-down is applying structured retrospective prompts. This approach enforces a blameless culture while applying five-axis scoring and run-ID correlation to preserve collaboration data.

Can I use this blameless retro skill for multi-agent governance without exposing PII?▼

Yes, you can use this skill for multi-agent governance without exposing PII. Blameless culture enforcement applies strict prompts and storage rules that prevent individual attribution and PII exposure, ensuring safe honest reporting.

Related Skills

Explore More Software Engineering Skills→
appautomatonappautomaton
130

collaborating-with-claude

Coordinate with Claude Code via a bridge workflow for prototyping, debugging, and code review.

Official
Intermediate
heroveherove
26

sleep-analyzer

Analyze sleep data to compute quality scores and generate improvement recommendations.

Community
Advanced
Corvalis-LLCCorvalis-LLC
6

auto-compliance

Enforce consistent privacy and compliance handling across user data interactions.

Official
Advanced
mcauduro0mcauduro0
1

scrum-master

Analyze sprint data with Monte Carlo simulations to forecast velocity and team health.

Community
Advanced