Unverified50% confidenceFactExact time
Appending rules to the system prompt when the model makes a mistake makes the Agent perform worse, because the system prompt is always resident, and every added character is fed into the model again at every reasoning step, continuously consuming attention and token budget.
1
Sources
50%
Confidence
Long-term
Relevance
10/24/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedOverly verbose, contradictory, or ambiguous system prompts cause the model's attention to become scattered, resulting in over-thinking or instruction-following failures; streamlining agents.md can improve the determinism of instruction following.79% similarUnverifiedBecause large models hallucinate and produce unstable outputs, agents in real-world engineering rely more on rules for execution, with most rules configured through prompts.77% similarUnverifiedVague instructions (such as "handle my inbox") cause the model to produce a different interpretation on each run, increasing unpredictability.75% similarUnverifiedFor complex creative tasks, cramming all requirements into a single prompt tends to yield poor results; providing context-relevant feedback in stages allows the model to focus on specific issues in each iteration.75% similarUnverifiedContext compression causes previously injected system prompts and custom skill instructions to be diluted or even lost, leading to model behavior drift that requires reloading.75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1093654API
curl https://kongchang.com/api/v1/knowledge/claims/1093654MCP
get_claim(id=1093654)