synthos

Autonomously scans skill trees and logs to detect and fix gaps without user instruction.

6|1|Updated May 11, 2026
One-click install
npx skills add https://github.com/yakeworld/Synthos --skill synthos-yakeworld
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: synthos
Source: https://github.com/yakeworld/Synthos/tree/main/skills/private/extended/research-tools/synthos-akne-bridge/synthos
Command: npx skills add https://github.com/yakeworld/Synthos --skill synthos-yakeworld

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Autonomous agents often stall by waiting for user instructions during idle cycles, after task failures, or between evolution rounds. This Skill enforces proactive behavior: it scans the skill tree and logs, identifies concrete gaps (such as missing verification checklists or constraint rules), and fixes them with traceable, source-backed output. ## Core Features & Use Cases - Proactive Exploration: Triggered by idle cycles, task failures, or evolution rounds, it autonomously scans the skill tree and system state instead of waiting for instructions. - Traceable Fixes: Every assertion and fix includes verifiable provenance (file paths, log line numbers) and modification diffs, with credentials read only from environment variables. - Boundary Enforcement: Detects functional overlap with other synthos-akne-bridge skills and merges to keep atomic boundaries clean. - Use Case: After an evolution round ends with no new tasks, the Skill scans the skill tree, finds that a content skill lacks a constraint-rules section, derives rules from that file's existing pitfalls and verification content, and applies the fix with a full diff and source references. ## Quick Start Trigger the synthos skill during an idle cycle to scan the skill tree, find a concrete gap, and autonomously fix it with traceable sources.

Frequently Asked Questions about synthos

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I make an AI agent act autonomously without user instructions?

Use trigger events such as idle cycles, task failures, or evolution rounds to drive autonomous scanning of the skill tree and logs. The agent must identify concrete gaps and execute fixes immediately rather than outputting a waiting-for-user response.

How to ensure agent outputs are traceable and verifiable?

Attach verifiable provenance to every assertion, such as file paths or log line numbers, and include modification diffs for any fix. Each source reference must be independently checkable, for example by confirming the file exists or locating the log line.

Can credentials be hardcoded in skill instructions or commands?

No. Credentials must always be read from environment variables, never written as plaintext in skill content or commands. Outputs are validated by grepping for credential patterns to confirm no hardcoded secrets exist.

What happens when two skills have overlapping functionality?

Overlapping skills must be merged immediately to preserve clear atomic boundaries. The skill checks whether its actions overlap with other skills in the same bridge and merges rather than duplicating functionality.

Why does the agent respond with waiting for user instructions?

That response indicates an autonomy lapse violating the proactive-exploration principle. The fix is to diagnose the lapse, return to scanning the skill tree for gaps or failed tasks, and execute a discovery action with source-backed output.