history-explorer

Aggregates decision-log entries into skill co-occurrence and meta-skill usage statistics as JSON.

6.8k|540|Updated May 6, 2026
One-click install
npx skills add https://github.com/opensquilla/opensquilla --skill history-explorer
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: history-explorer
Source: https://github.com/opensquilla/opensquilla/tree/main/src/opensquilla/skills/bundled/history-explorer
Command: npx skills add https://github.com/opensquilla/opensquilla --skill history-explorer

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

It answers questions about past agent behavior—such as which skills were used most often or which skills co-occur—by reading the per-turn DecisionEntry JSONL logs, without requiring manual log inspection.

Core Features & Use Cases

  • Co-occurrence Analysis: Aggregates DecisionEntry.skills_invoked over a configurable time window to surface frequently co-occurring skill chains.
  • Meta-Skill Usage Stats: Joins log data with bundled and managed meta-skill definitions to report meta-skill usage frequencies.
  • Router Fixture Corpus: Lists the router test-fixture corpus for downstream analysis.
  • Use Case: Ask "which skills did I use most this week?" and receive a JSON summary that a downstream LLM step (such as meta-skill-creator's harvest phase) can consume directly.

Quick Start

Ask the agent to show which skills co-occurred most often over the past 30 days using the history explorer.

Frequently Asked Questions about history-explorer

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

FAQPage Schema
How do I analyze which skills I used most in OpenSquilla?

Run the history explorer with a query and a window-days parameter to aggregate DecisionEntry logs into co-occurrence and meta-usage statistics. The script reads JSONL decision logs and prints a JSON summary to stdout.

How to query skill co-occurrence patterns from decision logs?

Invoke explore.py with --include co_occurrences and --window-days set to your desired range. The aggregator counts how often skills appear together in DecisionEntry.skills_invoked and returns the top-k pairs as JSON.

Where does the history explorer read decision logs from?

It resolves the log directory in order: the --log-dir argument, the OPENSQUILLA_LOG_DIR environment variable, OPENSQUILLA_STATE_DIR/logs, or ~/.opensquilla/logs. Tilde paths are expanded at runtime.

What happens when no decision log history exists?

The script returns an empty result with a placeholder string stating that no history is available and downstream consumers should rely on user intent only. It does not fail or raise an error.

Can the history explorer run outside a source checkout?

Yes, it works in wheel installs because it derives the package root from its own file location. Router fixtures are returned as an empty list when the test-fixture directory is absent.