Agent Insights

Scan a repository to surface actionable findings about coding agent performance, each with a targeted action recommendation.

What It Does

This plugin analyzes five data sources in parallel — source code, git history, GitHub PRs/CI, agent conversation logs, and agent configuration files — then cross-references and synthesizes the results into:

• Overall Score: A one-word assessment of the repo's agent-readiness — blocked, constrained, or productive
• Executive Summary: A short markdown paragraph giving a quick read on the state of the report before drilling into findings
• Findings: Specific, evidence-backed issues affecting agent performance, categorized by the APEX taxonomy
• Actions: Each finding carries an inline action — a targeted fix with rock-sized effort estimates (pebble/stone/rock/boulder) and Tessl product awareness (recommending plugins when appropriate)
• HTML Report: A self-contained, shareable report with filtering, search, and evidence provenance trails

Quick Start

run-insight-scan

This orchestrates all five analyzers and the synthesis step, producing output in .tessl-insights-poc/ at the repo root.

Prerequisites

• A git checkout of the target repository
• gh CLI installed and authenticated (for GitHub data analysis)
• Agent conversation logs on the machine (optional but recommended — from Cursor, Claude Code, etc.)

Multi-Developer Workflow

The scan results are designed to be committed to git so the whole team can contribute and benefit.

Initial scan (first developer)

1. Run run-insight-scan to analyze the repo
2. Review .tessl-insights-poc/report.html
3. Commit .tessl-insights-poc/ to git

Additional contributions (other developers)

1. Pull the latest .tessl-insights-poc/ from git
2. Run contribute-agent-logs — this analyzes your local agent logs and re-runs synthesis
3. Commit the updated .tessl-insights-poc/

Each developer's agent logs are saved separately (by username and date), so contributions are additive and don't overwrite each other. The synthesis step merges all available data on every run.

Output Structure

.tessl-insights-poc/
├── findings.json              # Synthesized findings with inline actions
├── report.html                # Self-contained HTML report
└── reports/                   # Raw analyzer outputs
    ├── source-code.json
    ├── git-history.json
    ├── github-data.json
    ├── agent-context.json
    └── agent-logs/
        ├── joe-20260408.json
        └── alice-20260410.json

Skills