Skip to content
Navigation menu
Search
Powered by Algolia
Search
Log in
Create account
DEV Community
Close
AI Agent Harness Engineering Benchmarks Series' Articles
Back to tappers's Series
I Benchmarked 41 AI Agent Harness Patterns in Isolated VMs: How to Cut Tokens by 90% Without Losing Accuracy
tappers
tappers
tappers
Follow
Oct 31
I Benchmarked 41 AI Agent Harness Patterns in Isolated VMs: How to Cut Tokens by 90% Without Losing Accuracy
#
ai
#
llm
#
agents
#
architecture
Comments
Add Comment
22 min read
Moderation
How to Benchmark AI Agents Without Eval Cheating or API Key Leaks: Designing a Proxmox & systemd Sandbox
tappers
tappers
tappers
Follow
Oct 31
How to Benchmark AI Agents Without Eval Cheating or API Key Leaks: Designing a Proxmox & systemd Sandbox
#
ai
#
security
#
proxmox
#
sandbox
Comments
1
comment
14 min read
Moderation
Stop Your Coding Agent From Stepping on the Same Rake Twice: Cold-Start Exploration vs. a 5-Item MEMORY.md
tappers
tappers
tappers
Follow
Oct 31
Stop Your Coding Agent From Stepping on the Same Rake Twice: Cold-Start Exploration vs. a 5-Item MEMORY.md
#
ai
#
llm
#
agents
#
harness
Comments
Add Comment
8 min read
Moderation
Why Feeding Full Dependency Files to Your Coding Agent Hurts Accuracy: Full Source vs. Tree-sitter Interface Skeletons
tappers
tappers
tappers
Follow
Nov 1
Why Feeding Full Dependency Files to Your Coding Agent Hurts Accuracy: Full Source vs. Tree-sitter Interface Skeletons
#
ai
#
llm
#
agents
#
contextengineering
Comments
Add Comment
7 min read
Moderation
Stop Dumping Raw Pytest Logs Into Your Coding Agent: How a Structured JSON Test Filter Cuts Input Tokens by 94.1%
tappers
tappers
tappers
Follow
Nov 1
Stop Dumping Raw Pytest Logs Into Your Coding Agent: How a Structured JSON Test Filter Cuts Input Tokens by 94.1%
#
ai
#
llm
#
agents
#
testing
Comments
Add Comment
6 min read
Moderation
Why Keeping Failed Code Attempts in Chat History Poisons Your Coding Agent: Error-Trace Pruning vs. Accumulated History
tappers
tappers
tappers
Follow
Nov 1
Why Keeping Failed Code Attempts in Chat History Poisons Your Coding Agent: Error-Trace Pruning vs. Accumulated History
#
ai
#
llm
#
agents
#
contextengineering
Comments
Add Comment
6 min read
Moderation
The Hidden "Schema Tax" of MCP Tools: Dumping 32 Tool Schemas vs. On-Demand Tool Search (-92.7% Input Tokens)
tappers
tappers
tappers
Follow
Nov 1
The Hidden "Schema Tax" of MCP Tools: Dumping 32 Tool Schemas vs. On-Demand Tool Search (-92.7% Input Tokens)
#
ai
#
llm
#
agents
#
mcp
Comments
2
comments
6 min read
Moderation
How Save-Game Context Compaction Fixes Spec Drift in Multi-Step Coding Agents (-91.9% Input Tokens)
tappers
tappers
tappers
Follow
Nov 1
How Save-Game Context Compaction Fixes Spec Drift in Multi-Step Coding Agents (-91.9% Input Tokens)
#
ai
#
llm
#
agents
#
contextengineering
Comments
Add Comment
6 min read
Moderation
Benchmarking the Claude Code /advisor Pattern: Solo Executor Loops vs. Upfront Advisor Blueprint Routing
tappers
tappers
tappers
Follow
Nov 1
Benchmarking the Claude Code /advisor Pattern: Solo Executor Loops vs. Upfront Advisor Blueprint Routing
#
ai
#
llm
#
agents
#
architecture
Comments
Add Comment
7 min read
Moderation
Single Loops vs. Directed Graphs for AI Coding Agents: Cutting Input Tokens by 74.4% Without Losing Accuracy
tappers
tappers
tappers
Follow
Nov 16
Single Loops vs. Directed Graphs for AI Coding Agents: Cutting Input Tokens by 74.4% Without Losing Accuracy
#
ai
#
harness
#
architecture
#
agenticai
Comments
Add Comment
6 min read
Moderation
Why Role-Based Subagents Slow Down Coding Agents: Subagent Folding + Progressive Skill Disclosure (2.66x Faster)
tappers
tappers
tappers
Follow
Nov 16
Why Role-Based Subagents Slow Down Coding Agents: Subagent Folding + Progressive Skill Disclosure (2.66x Faster)
#
ai
#
harness
#
contextengineering
#
multiagent
Comments
Add Comment
6 min read
Moderation
Why AI Agents Write Weak Tests for Their Own Code: Single-Agent TDD vs. Two-Agent Blind TDD (0% vs. 100% Mutant Kill Rate)
tappers
tappers
tappers
Follow
Nov 17
Why AI Agents Write Weak Tests for Their Own Code: Single-Agent TDD vs. Two-Agent Blind TDD (0% vs. 100% Mutant Kill Rate)
#
ai
#
harness
#
tdd
#
testing
Comments
Add Comment
5 min read
Moderation
We're a place where coders share, stay up-to-date and grow their careers.
Log in
Create account