skills.rest
brandonavant
brandonavantCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

case-study-capture

Log manual observations with predefined categories to a JSONL case study file.

2|1|Updated Mar 15, 2026
Search Tags:#logging#claude.md#audit-trail#harness#case-study#context-summary#manual-observation
One-click install
npx skills add https://github.com/brandonavant/harness-engineering-guide --skill case-study-capture
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: case-study-capture
Source: https://github.com/brandonavant/harness-engineering-guide/tree/main/case-study-harness/claude/skills/case-study-capture
Command: npx skills add https://github.com/brandonavant/harness-engineering-guide --skill case-study-capture

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Log manual observations about the current session to capture insights that automatic hooks miss, enabling richer case studies and traceable decisions.

Core Features & Use Cases

  • Logs a single observation to the harness case study log.
  • Classifies the observation into one of: successful_pattern, human_override, context_architecture, friction, or other.
  • Persists the entry to case-study-harness/data/manual-observations.jsonl for auditing and analysis.

Quick Start

Describe the observation and run the script with the category, your description, and a concise context summary.

Frequently Asked Questions about case-study-capture

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

FAQPage Schema
How do I log manual observations to capture session insights that automatic hooks miss?▼

To log manual session observations, you run the provided Python script with a predefined category, your description, and a context summary. The script persists the entry as a JSONL file for auditing and analysis.

What categories can I use to classify friction points or context architecture notes in an audit trail?▼

You can classify audit trail observations into five predefined categories: successful_pattern, human_override, context_architecture, friction, or other. This ensures consistent manual logging across harness runs.

When do I need to manually log context summaries instead of relying on automated session tracking?▼

You need manual context summaries when recording human overrides, friction points, or context-architecture notes that automated session tracking cannot detect. This captures traceable decisions and richer case studies across harness runs.

Where are manual session observations stored for case study analysis?▼

Manual session observations are persisted to the case-study-harness/data/manual-observations.jsonl file. This JSONL format ensures append-only logging for reliable auditing and case study analysis.

Do I need any dependencies to run the Python script for logging session observations?▼

No external dependencies are required to run the Python script for logging session observations. You only need the script itself along with your category, description, and context summary inputs.

Related Skills

Explore More Content & Communication Skills→
kandengkandeng
2

hello-python

Execute a pre-built Python 3 script to generate personalized greetings.

Community
Basic
brandonavantbrandonavant
2

case-study_synthesize

Synthesize observation data into a case study draft and improvement plan.

Community
Advanced
dnvitidnviti
10

test-review

Run automated and manual tests for tasks in 'to-test' status.

Community
Advanced
outfitter-devoutfitter-dev
6

trails

Log research findings and create structured session handoff notes.

Official
Intermediate