AI Briefing#prompt-engineering
The latest AI and developer news about #prompt-engineering, with the original source and a short summary.
The latest AI and developer news about #prompt-engineering, with the original source and a short summary.
The method uses a repeatable cycle of proposing, testing, and learning to fix agent behaviors without breaking existing functionality.
Harness size was halved due to improved model performance, but verification mechanisms must be maintained.
In 2,160 simulated outage scenarios, Claude Sonnet 5.5 and GPT-6.1 Sol emerged as leaders 1.7 times more often than Gemini 3.1 Pro, while agents listed first in the system prompt issued plans three times more frequently.
The feature uses signals like label changes and PR merges to propose prompt refinements, which users must review and approve before application.
Anthropic advises defaulting to one model and context window, noting that tool choice degrades past 30-50 definitions and that a single loop ran 2x faster than multi-agent harnesses.
Kokon 0.10 introduces a 'Prompt+' feature that uses the existing Qwen3-VL 8B encoder to expand prompts, adding roughly 2 seconds of processing time on an RTX 5070 Ti, alongside Windows-only support for custom Qwen-Image 2.1 LoRAs.
The method uses GPT-4.1 to extract behavioral guidelines from agent trajectories, improving reliability and accuracy without weight updates or human annotation.
NHN Cloud applied the AI harness 'Silver Bullet' to team workflows, reducing incident root-cause analysis time to 4–10 minutes.
The open-source Terse plugin for Claude Code enforces 22 writing rules to eliminate verbose 'Claude-speak,' with benchmarks showing word count reductions of 46-54% and cost savings of 32-50%.
Prevent unnecessary context accumulation and improve exploration efficiency through progressive disclosure and sub-agent utilization
Hardcoded AI code was separated into CSV rule sets to enable maintenance by non-developers.
Claude Opus 5.5 defaults to 'medium' effort and requires new handling for thinking blocks and safety refusals compared to Opus 5.
The new commands automate eval design and iterative improvements, achieving up to 98.9% accuracy at one-fifth the cost in internal benchmarks.
A firsthand account of spending a month refining an AI harness to solve the problem of repetitive initialization in every AI session.
Reflects API structural changes including a 30% output speed improvement over Opus 5, forced thinking activation, and default effort level changes.
The open-source library tool-prune has been released, reducing prompt tokens by 92% through pre-filtering of tool schemas.
The OpenAI Codex team released a guide to removing unnecessary instructions and managing context budgets in line with GPT-6 Astra's performance improvements.
In experiments based on GPT-5.6, instructions for verification skills such as TDD and Formal Methods increased costs without improving accuracy.
An open-source skill has been released that requires AI coding agents to write detailed briefs before coding.
Anthropic has removed over 80% of the system prompts for its Claude 5 models and unveiled six principles of judgment-centric context engineering without performance degradation.
A Goal-Driven prompt has been released, where a master agent verifies completion criteria to re-run sub-agents.
System prompts are high-level instructions that define an AI model's role, tone, and guardrails to ensure consistent behavior.
Vercel has released a single file, 'design.md', to ensure external AI agents adhere to its design system.
Olive Young launched a dedicated ChatGPT plugin and revealed a prompt optimization strategy where PMs and developers collaborated to reduce LLM parameter misinterpretation.