← Back to Skills Marketplace
harrylabsj

Insight Tracker

by haidong · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
277
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install insight-tracker
Description
Track and manage insights, patterns, and observations discovered during OpenClaw sessions. Use when the user wants to record, categorize, retrieve, or analyz...
README (SKILL.md)

Insight Tracker

Overview

This skill provides structured insight management for OpenClaw workflows. It captures observations, patterns, and learnings that emerge during sessions, making them searchable, referenceable, and actionable.

When to Use

Use this skill when:

  • A user mentions "note this down" or "remember this insight"
  • Patterns emerge across multiple sessions that should be tracked
  • Research or analysis reveals findings worth preserving
  • The user asks to "track" or "record" something for future reference
  • Cross-referencing insights with existing knowledge is needed

Core Concepts

Insight

An insight is a discrete observation, pattern, or learning with the following attributes:

  • Content: The actual observation or finding
  • Source: Where it came from (session, research, user input)
  • Tags: Categories for organization
  • Priority: Importance level (high/medium/low)
  • Status: Current state (active/validated/archived)
  • Created: Timestamp
  • References: Links to related insights or knowledge

Tags

Standard tags for categorization:

  • pattern: Recurring behaviors or structures
  • learning: New understanding or skill acquired
  • decision: Choices made and their rationale
  • risk: Potential issues or concerns
  • opportunity: Potential improvements or wins
  • user-preference: User-specific preferences
  • technical: Technical findings or constraints
  • process: Workflow or methodological insights

Input

Accepts insights in various formats:

  • Direct text input
  • Session transcript excerpts
  • Memory file references
  • Research findings

Output

Produces:

  • Dated insight records (Markdown)
  • Tagged insight summaries
  • Cross-reference reports
  • Insight search results

Workflow

Recording an Insight

  1. Capture: Extract the core observation
  2. Tag: Apply relevant category tags
  3. Prioritize: Mark importance level
  4. Link: Connect to related insights
  5. Store: Save to dated record

Retrieving Insights

  1. Search: By tag, keyword, date, or priority
  2. Filter: Narrow by status or category
  3. Present: Show matching insights with context

Analyzing Insights

  1. Cluster: Group related insights
  2. Trend: Identify patterns over time
  3. Validate: Check against new evidence
  4. Archive: Mark outdated insights

Commands

Add Insight

insight add "Content of the insight" --tags pattern,learning --priority high

List Insights

insight list --tag pattern --since 2024-01-01

Search Insights

insight search "keyword" --priority high

Show Insight

insight show \x3Cinsight-id>

Archive Insight

insight archive \x3Cinsight-id>

Output Format

Dated Insight Record

# Insights - YYYY-MM-DD

## New Insights

### INS-001: Title
- **Content**: The insight content
- **Source**: Session/user/research
- **Tags**: pattern, learning
- **Priority**: high
- **Status**: active
- **Created**: 2024-01-15T10:30:00Z
- **References**: INS-003, knowledge-distillation-2024-01-10

## Summary
- Total insights: 5
- High priority: 2
- New tags: user-preference

Quality Rules

  • Be specific: avoid vague generalizations
  • Include source: always note where insight came from
  • Tag consistently: use standard tags
  • Link related insights: build knowledge networks
  • Review regularly: archive outdated insights
  • Prioritize honestly: not everything is high priority

Good Trigger Examples

  • "Track this insight: users prefer X over Y"
  • "Note down that we discovered a pattern in Z"
  • "Search for insights about deployment issues"
  • "Show me all high priority insights from last week"
  • "Archive insight INS-005, it's no longer relevant"

Resources

references/

  • references/tag-taxonomy.md: Full tag definitions and usage guidelines
  • references/output-examples.md: Sample insight records
Usage Guidance
This skill appears to be a simple local insight tracker implemented with shell scripts and local markdown files, but there are a few mismatches and sloppy bits you should check before installing: - SKILL.md documents a single `insight` CLI and commands like `insight search` and `insight archive`, but the repo only contains separate scripts (add-insight.sh, list-insights.sh) and lacks a search/archive implementation and an 'insight' wrapper. Confirm whether your agent runtime provides an `insight` command or whether the skill author intended to include a wrapper. - test.sh appears to try to use a temporary data directory but does not export INSIGHT_DATA_DIR when calling add-insight.sh, so running tests may write files under the skill folder's data/ by default. Review and, if needed, run tests in a safe sandbox to avoid polluting directories. - The scripts write to a local data directory (default: <skill>/data) or to the directory you provide via INSIGHT_DATA_DIR. Decide where you want those files stored and ensure file permissions and backups meet your requirements. - The add-insight script interpolates user-provided content into YAML and markdown files; if you plan to feed untrusted content into it (e.g., automated transcripts), review sanitization/quoting to avoid malformed frontmatter or unexpected injections. If you are comfortable with local file-based storage and can tolerate/fix the documentation mismatch and test bug, the skill is probably usable. If you expect a unified CLI (insight) or remote storage/search capabilities, ask the author for a corrected release or a README explaining how the provided scripts map to the documented commands.
Capability Analysis
Type: OpenClaw Skill Name: insight-tracker Version: 1.0.0 The skill contains a shell injection vulnerability in `scripts/add-insight.sh` because the `$CONTENT` variable is expanded inside an unquoted heredoc (`<< EOF`), allowing for arbitrary command execution if the input contains shell metacharacters (e.g., backticks or `$()`). While the skill's logic for tracking and categorizing insights appears legitimate and aligned with its stated purpose, the lack of input sanitization in the bash scripts poses a significant security risk.
Capability Assessment
Purpose & Capability
The skill claims to provide insight recording, tagging, searching, and reporting and the repository contains shell scripts that implement adding and listing insights stored as local files. However, the SKILL.md shows a single 'insight' CLI (e.g., `insight add`, `insight list`, `insight search`) while the repo provides separate scripts (add-insight.sh, list-insights.sh) and lacks an 'insight' wrapper or a search implementation matching the documented commands. This is likely documentation mismatch or incomplete wiring rather than malicious, but it is an inconsistency.
Instruction Scope
SKILL.md describes accepting session transcripts, memory file references, and cross-referencing with existing knowledge, but the shipped scripts operate on local markdown files in a data directory and do not implement search/show/archive features as documented (no explicit 'search' or 'archive' script). The SKILL.md commands could therefore mislead an agent into trying to run nonexistent commands. The instructions do not attempt to read unrelated system files or network endpoints.
Install Mechanism
No install spec or external downloads are present. The skill is instruction-and-script-only; files are included in the package and no external code is fetched. This is low-risk from an install-mechanism perspective.
Credentials
No credentials or sensitive environment variables are required. Scripts optionally honor INSIGHT_DATA_DIR to choose where data files are stored; that is proportional to the stated purpose. There are no requests for unrelated credentials or config paths.
Persistence & Privilege
The skill does not request permanent/always-on inclusion, does not modify other skills, and only writes data into its own data directory (or a directory set by INSIGHT_DATA_DIR). No elevated privileges are requested.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install insight-tracker
  3. After installation, invoke the skill by name or use /insight-tracker
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release.
Metadata
Slug insight-tracker
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is Insight Tracker?

Track and manage insights, patterns, and observations discovered during OpenClaw sessions. Use when the user wants to record, categorize, retrieve, or analyz... It is an AI Agent Skill for Claude Code / OpenClaw, with 277 downloads so far.

How do I install Insight Tracker?

Run "/install insight-tracker" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Insight Tracker free?

Yes, Insight Tracker is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Insight Tracker support?

Insight Tracker is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Insight Tracker?

It is built and maintained by haidong (@harrylabsj); the current version is v1.0.0.

💬 Comments