← All concepts

system prompt engineering

5 articles · 15 co-occurring · 0 contradictions · 2 briefs

The length constraint added to system prompts directly caused coding quality regression. This is a concrete example of how prompt design decisions impact output quality.

Related concepts
prompt engineering 4 tool integration patterns 2 transparency in ai systems 1 tool use architecture 1 state management 1 session persistence 1 retrieval augmented generation 1 prompt optimization 1 model selection strategy 1 model context protocol 1 inference parameter configuration 1 context window optimization 1 context window management 1 context window constraints 1 context caching strategy 1
Signal history
2026-W15
8
Evidence chain (5 articles, showing 5)
Context Engineering Explained: The Post-Prompt-Engineering 2026 Beginner's Guide UK Business Owners Have Been Asking For example_of

Explicitly mentioned as component of context engineering practice in BraivIQ offerings

Claude Code's Performance Issues: Anthropic Admits to Changes | Raffi Krikorian posted on the topic | LinkedIn example_of

The length constraint added to system prompts directly caused coding quality regression. This is a concrete example of how prompt design decisions impact output quality.

@steipete: acpx 0.6.0 is out. (control codex/claude via agents) example_of

'Claude system-prompt controls' is explicit system prompt management, a core context engineering lever.

@emollick: Reading one of the viral "anti-slop" markdown files for agents and realizing ... example_of

The 'viral markdown files for agents' are system prompt/context engineering artifacts; Mollick is identifying a failure mode in how these are generated (by AI rather than humans)

@paoloanzn: anthropic is now blocking THEIR OWN cli tool if they suspect you're using it ... example_of

The blocking specifically targets system prompt modification, a core context engineering technique

query this concept
$ db.articles("system-prompt-engineering")
$ db.cooccurrence("system-prompt-engineering")
$ db.contradictions("system-prompt-engineering")