[FEATURE]: System prompt over specification audit and mitigation

[micuintus]

Feature hasn't been suggested before.

• I have verified this feature I'm about to request hasn't been suggested before.

Describe the enhancement you want to request

This is PR1 of #14398

Overspecification Hurts Performance

Two landmark studies prove that excessive specification in prompts degrades LLM performance:

Study	Finding	Evidence
Tang et al. (2025)	Over-prompting degrades performance	Performance peaks at 5-20 examples, then declines
Zi et al. (2025)	More detail ≠ better code generation	Excessive constraints create cognitive overload

Implication: Too few constraints → model ignores them. Too many → performance degrades.
Optimal: 5-7 high-impact constraints per category.

PR-1: Overspecification Audit & Reduction

Priority Score: 9.6/10 (HIGH impact, VERY HIGH evidence, LOW effort, ZERO overspec risk)

What: Audit all prompts and reduce constraints to 5-7 per category. Remove redundant instructions, verbose explanations, and duplicate content. Target <1500 tokens per prompt.

Why: Research proves excessive constraints degrade performance. Tang et al. (2025) found performance peaks at 5-20 examples then declines. Zi et al. (2025) confirmed excessive constraints create cognitive overload.

Evidence:

• Tang et al. (2025): Performance peaks at 5-20 examples, then declines
• Zi et al. (2025): Excessive constraints create cognitive overload
• Chen et al. (2025): Promptware anti-patterns identified

Tasks:

• Audit all prompt files in packages/opencode/src/session/prompt/*.txt
• Count MUST/NEVER rules per category (target: ≤7)
• Count examples per file (target: ≤5)
• Measure token count per file (target: <1500)
• Remove redundant/duplicate instructions
• Create validation script scripts/audit-overspecification.ts
• Add to CI pipeline

Files: packages/opencode/src/session/prompt/*.txt, scripts/audit-overspecification.ts

[github-actions]

This issue doesn't fully meet our contributing guidelines.

What needs to be fixed:

• The description appears to be AI-generated (structured research tables, "Priority Score: 9.6/10", academic citation formatting, etc.). We don't accept AI-generated walls of text as issue content — please rewrite this in your own words describing the problem and what you want changed.
• This issue is framed as a sub-task ("PR1 of [FEATURE]: Evidence based refurbishment of openCode's System Prompt (Architecture) #14398") rather than a standalone feature request. Issues should stand on their own; if this is part of a larger initiative, please describe the specific, self-contained change being requested here without relying on another issue as context.

Please edit this issue to address the above within 2 hours, or it will be automatically closed.

---

This issue might be a duplicate of existing issues. Please check:

• [FEATURE]: Evidence based refurbishment of openCode's System Prompt (Architecture) #14398: Same author, same topic — system prompt architecture improvement (this issue references it directly as its parent)
• [FEATURE]: Optimize prompts using principle constraints and DSL compression #6249: Very similar request — auditing and optimizing prompts to reduce verbosity and token count

If you believe this was flagged incorrectly, please let a maintainer know.

[micuintus]

@thdxr Compliance should be met?

[github-actions]

This issue has been automatically closed because it was not updated to meet our contributing guidelines within the 2-hour window.

Feel free to open a new issue that follows our issue templates.