Self Improving To Expertpack

Convert Self-Improving Agent learnings into a structured ExpertPack. Migrates the .learnings/ directory (LEARNINGS.md, ERRORS.md, FEATURE_REQUESTS.md) and an...

This skill appears coherent and limited to converting .learnings/ content into an ExpertPack; it reads files under the workspace you point it at and writes the resulting pack to the output directory. Before running: (1) run it on a copy or a non-sensitive workspace to confirm behavior; (2) note that 'secret stripping' is regex-based (common patterns like sk-*, ghp_*, xoxb- and simple key:value patterns are redacted) but may miss nonstandard secrets—inspect outputs for leftover sensitive data; (3) check the generated files (overview.md, manifest.yaml, content files) before publishing or committing them; (4) ensure you pass a workspace path you control (the script will read workspace files you specify). If you need network or cloud integration, this skill does not request credentials and does not perform network calls according to the provided code.

Type: OpenClaw Skill Name: self-improving-to-expertpack Version: 1.0.1 The skill bundle is a utility for migrating agent learning logs into a structured 'ExpertPack' format. The core logic in scripts/convert.py is a standard file-processing script that parses markdown, categorizes content, and generates new files; notably, it includes proactive security features such as regex-based redaction of API keys and tokens (SECRET_PATTERNS) to prevent accidental exposure of credentials during the migration process.