Reflect
Self-improvement through conversation analysis. Extracts learnings from corrections and success patterns, permanently encoding them into agent definitions. Philosophy - Correct once, never again.
Why this rating
Deterministic checks triggered by the tool capabilities and evidence.
- Locality: Local
Stores state under `~/.reflect` and writes outputs to local project paths.
- Data access: Sensitive
Reviews conversation content and project/agent instruction files.
- Action surface: Execute
Declares `Bash`/`Write`/`Edit` tools and applies approved changes.
Best practices
Follow these steps to reduce risk when using this skill.
- Require explicit approval before applying any proposed file edits or commits.
- Protect `~/.reflect` state files and avoid storing secrets in reflection logs.
- Review generated diffs carefully to prevent unsafe instruction drift.
Evidence links
Public sources backing the indicator assignments.
Always be careful when navigating away from the website.
Max-risk rule
If any capability reaches a higher level, the entire indicator level bumps up to keep ratings deterministic and easy to scan.