skills.rest
Dozyboy
DozyboyCOMMUNITY
Explore all skills by this author →
GitHub

hs:insights

Analyzes harness telemetry JSONL logs to surface skill usage patterns and advisory optimization suggestions.

Updated Jul 19, 2026
Search Tags:#skill usage#workflow analysis#telemetry#read-only#usage analytics#jsonl#gate monitoring
One-click install
npx skills add https://github.com/Dozyboy/VSF --skill hs-insights-dozyboy
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: hs:insights
Source: https://github.com/Dozyboy/VSF/tree/main/Day2_VSF/Demo1/harness/plugins/hs/disabled-skills/insights
Command: npx skills add https://github.com/Dozyboy/VSF --skill hs-insights-dozyboy

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Teams running a harness accumulate telemetry data but lack visibility into how skills are actually used, which workflows chain together, and where gates block or get bypassed. This Skill turns raw telemetry JSONL logs into a narrated, read-only insights report with actionable suggestions. ## Core Features & Use Cases - Usage analysis: Identifies hot skills and never-invoked owned skills as trim or merge candidates, without ever auto-removing anything. - Workflow and gate lenses: Compares observed workflow chains against declared skill-chains.yaml, and reports gate pass/block/advisory trends with top block reasons. - Session shape reporting: Computes duration p50/p90, tool mix, and files-modified totals from sessions.jsonl to show where time actually goes. - Use Case: Run a 30-day review before a sprint retro to discover that three owned skills were never invoked and that a rising gate_advisory count signals receipt gaps that would block on remote. ## Quick Start Ask the assistant to analyze the last 30 days of harness telemetry and summarize hot skills, unused skills, and gate-block patterns.

Frequently Asked Questions about hs:insights

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

FAQPage Schema
How do I analyze skill usage from telemetry logs?▼

Run analyze_telemetry.py with --lens all and --days 30 against the harness telemetry JSONL sinks. The report shows hot skills, never-invoked owned skills, workflow chains, and gate patterns, with a NOT measured block listing what telemetry does not capture.

What telemetry lenses are available for usage analysis?▼

Available lenses are workflow, skill_usage, session, gate, perf_trend, and all. Each lens reads a specific JSONL sink, such as sessions.jsonl for duration percentiles or the gate trace for pass, block, and advisory counts.

Does the insights report modify my configuration or skills?▼

No, the analysis is strictly read-only. It runs analyze_telemetry.py and reads JSONL sinks without writing; every suggestion is advisory and acting on one goes through normal tools like /hs:setup or a backlog entry.

Why does the report suppress recommendations for some lenses?▼

Lenses marked gated: true fall below a low-volume threshold, meaning the data is too sparse to support reliable recommendations. The report shows raw counts only and explicitly notes the suppression to avoid over-reading a few data points.

What are the limitations of telemetry-based usage analysis?▼

Telemetry does not capture cost, correctness, or non-script wall-clock time, as listed in the NOT measured block. A clean metric should not be read as full coverage, and sparse data below the gating threshold yields counts only, not recommendations.

Related Skills

Explore More Software Engineering Skills→
OpenLAIROpenLAIR
1.1k

aris-meta-optimize

Analyzes ARIS usage logs to propose data-driven optimizations for SKILL.md prompts and workflow defaults.

Official
Advanced
sky770825sky770825

clawsec-suite

Monitor advisory feeds and verify installed AI skill integrity.

Community
Advanced
XGanQAQXGanQAQ
1

retrospective

Generate structured markdown sprint retrospectives from git history and plan data.

Community
Intermediate
firetiger-ossfiretiger-oss
1

firetiger-instrument

Automate OpenTelemetry instrumentation integration across Node.js, Python, Go, and Rust.

Official
Intermediate