Prompt file imported from autocloudarc-digital-services/ghcp-dotnet-calculator (.github/prompts/14.04-cates-token-efficiency-remediation.prompt.md). Copyright stays with the author.

Remediate CATES Token Efficiency

Objective

Reduce recurring configuration waste in the sample repository without starving the coding agent of security, scope, testing, or error-handling context.

Required Steps

1. Read cates-exercises/04-cates-token-efficiency-remediation.md and the Exercise 03 baseline report.
2. Triage TE findings by severity, recurring scope, and expected token impact.
3. Edit only coding-agent files under cates-exercises/workspace/sample-repository/.
4. Remove generic filler and exact duplication.
5. Replace unconditional verbosity and negative lists with one concise, conditional positive output contract.
6. Move inventory-specific examples to on-demand scope.
7. Replace unbounded includes with specific references or a compact layout.
8. Remove volatile current-state directives from always-loaded instructions.
9. Re-analyze with the original tokenizer and save reports/04-token-efficiency.json.

Safety Boundary

Do not remove prompt protection, least-privilege guidance, testing requirements, or failure escalation solely to lower tokens or raise the score.

Validation

Compare before and after by rule and loading scope. Report any useful context retained despite its token cost.