What sources are used for insights?

Insights are grouped by source: Pondering (background analysis) and Documentation Gaps (missing docs). Each item includes content, confidence, and generation time.

How are confidence scores used?

Scores are shown when they exceed the threshold. You can set min_confidence (e.g., --min-confidence 0.7) to tailor results.

What if there are no insights?

The system suggests actions to generate them: work on code to trigger pattern detection, make decisions to trigger pondering analysis, or use tools consistently to trigger workflow insights.

insights

npx machina-cli add skill ConaryLabs/Mira/insights --openclaw

Files (1)

SKILL.md

1.2 KB

Mira Insights

Current Insights (Live)

!mira tool insights '{"action":"insights","min_confidence":0.5}'

Instructions

Present the insights above grouped by source:

Pondering: Background analysis of your work patterns (stale goals, fragile code, revert clusters, untested hotspots)
Documentation Gaps: Missing docs that should be written

For each insight show:

The insight content
Confidence score (if above threshold)
When it was generated

If no insights exist, suggest actions that generate them:

Working on code (triggers pattern detection)
Making decisions (triggers pondering analysis)
Using tools consistently (triggers workflow insights)

Optional Filters

If user passes arguments, use mcp__mira__insights tool with filters:

--source pondering|doc_gap → insight_source parameter
--min-confidence 0.7 → min_confidence parameter

Source

git clone https://github.com/ConaryLabs/Mira/blob/main/plugin/skills/insights/SKILL.mdView on GitHub

Overview

Provides grouped, source-tagged insights about Mira's work patterns and background analysis. It helps you understand what patterns are emerging and what documentation is missing. If no insights exist, it suggests actions to generate them.

How This Skill Works

Mira fetches insights via the insights tool (action: insights, min_confidence threshold). Insights are grouped by source (Pondering and Documentation Gaps) and shown with content, confidence (when above threshold), and generation time. If no insights exist, the system proposes actions to generate them (e.g., working on code, making decisions, using tools consistently).

When to Use It

When you ask what Mira learned, want to see insights, or inquire about patterns noticed.
When you want Mira's background analysis or predictions.
When you need insights grouped by source (Pondering or Documentation Gaps) with confidence and timestamps.
When no insights exist and you need guidance on generating them.
When you want to filter results using --source pondering|doc_gap or --min-confidence to tailor insights.

Quick Start

Step 1: Trigger insights with the appropriate prompt or view Current Insights (Live).
Step 2: Review results grouped by source (Pondering, Documentation Gaps) with content and timestamps.
Step 3: Act on insights or adjust filters/min_confidence to refine results.

Best Practices

Review insights by source to distinguish background analysis from missing documentation.
Check confidence scores before acting; adjust min_confidence for precision.
Trigger pattern detection by actively working on code to refresh insights.
Use insights to identify hotspots: stale goals, fragile code, or untested areas.
Document gaps surfaced by insights to close knowledge gaps quickly.

Example Use Cases

After a sprint, Mira outputs pondering insights about stale goals and revert clusters.
A gap in API docs is surfaced under Documentation Gaps, prompting documentation work.
Insights are retrieved with min_confidence 0.7 to focus on high-confidence findings.
No insights exist; system suggests actions like coding, decision-making, or tool usage to generate them.
Insights guide refactoring decisions by highlighting unreproducible patterns and hotspots.

Frequently Asked Questions

Add this skill to your agents