MBTI Analyzer

Analyze a user's MBTI from authorized OpenClaw memory, session history, and workspace notes. Use when the user asks for MBTI analysis, personality inference...

This skill appears coherent for generating MBTI reports from your local OpenClaw history and workspace notes, but it reads potentially sensitive local data. Before installing or running it: 1) Confirm and explicitly authorize only the source categories you want analyzed (do not allow broad access by default). 2) Review discover_sources.py and ingest_all_content.py to verify they actually honor the declared exclusions (.env, credentials/*, identity/*) and do not read paths you consider sensitive. 3) Search the included Python files for network/networking calls (requests, urllib, http, socket), subprocess usage, or hard-coded external endpoints; if present, inspect what data they send and to whom. 4) Prefer running the skill in an isolated environment or on a copy of your workspace if you need to be cautious. 5) Disable quoting in the pipeline if you do not want any text excerpts included in the report. If you want higher assurance, request a full code review of the omitted files (build_evidence_pool.py, infer_mbti.py, render_report.py, mbti_common.py, discover/ingest scripts) to confirm there is no unexpected data exfiltration or filesystem access beyond the stated sources.

Type: OpenClaw Skill Name: mbti-analyzer Version: 0.4.0 The mbti-analyzer skill is a comprehensive tool designed to infer a user's MBTI profile by analyzing local OpenClaw session history, workspace notes (MEMORY.md), and task metadata. While the skill requires high-privilege access to sensitive local data (including SQLite databases and session logs in ~/.openclaw), the implementation includes several security-conscious features: it explicitly instructs the AI agent to seek user authorization before reading any sources, implements a 'pseudo-signal' filter to prevent the agent from misinterpreting system instructions as user personality traits, and defines clear exclusions for sensitive files like .env and credentials. The code (scripts/ingest_all_content.py, scripts/infer_mbti.py) performs all processing locally, and no evidence of data exfiltration, obfuscation, or unauthorized remote execution was found.