Self Review

Automatically evaluates and approves agent outputs based on clarity, conciseness, actionability, and structure using a rule-based system.

Self-Review Skill

Automatically review agent output quality before sending to user.

• Checks: clarity, conciseness, actionability, structure
• Simple rule-based engine (no API cost)
• Exit code 0 = approved, 1 = needs improvement

Usage

Pipe output to reviewer:

echo "Your response text" | node skills/self-review/index.js

Or integrate into agent pipeline (AGENTS.md step 6).

Configuration

Edit skills/self-review/index.js to adjust thresholds.

Advanced

For LLM-based review, see self-review-llm skill (separate package).

---

Author: dvinci达芬奇 (self-evolved) Version: 1.0.0 Tags: quality, automation, token-optimization