generalize-once

A rule for treating one reported problem as an example of a wider pattern. It asks the coding agent to find and fix related cases across the relevant code, configuration, tests, and documentation.

In plain words
What is it for?
Use when fixing repeated dead code, duplicated logic, stale documentation, missing validation, or similar patterns across a project.
Why use it?
It prevents repeated back-and-forth over obvious sibling problems and avoids leaving references or duplicate issues behind.

Cursor rule

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add rules/liuyihey/agent-engineering/generalize-once
Clone the repo
git clone --depth 1 https://github.com/LiuYihey/Agent-Engineering
Per session 340 This file is loaded in full into every session.
When invoked 340 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00340 $0.00340
Opus 5 $0.00170 $0.00170
Sonnet 5 $0.00068 $0.00068
Haiku 4.5 $0.00034 $0.00034

Measured 2d ago against content hash 9c11b4948aca, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

generalize-once scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

rules/generalize-once.mdc · 32 lines

What it actually says

Generalize Once, Fix Completely

When the user points out one problem, treat it as a class of problems. Deliver the full fix before they have to ask again for siblings.

Infer the real ask

  • What category is this? (dead code, redundant config, duplicated logic, stale docs, missing validation, ...)
  • What scope does it naturally cover? (same file, same module, same pattern across repo)
  • What would the user likely ask next if you only patch the single line they highlighted?

Answer those three, then fix that scope - not just the cited line.

One-pass workflow

  1. Search the pattern - grep / explore for the same smell elsewhere (sibling fields, parallel files, mirrored stubs, outdated docs).
  2. Fix source + consumers together - config, loader, API, UI, tests, docs that reference the same thing; don't leave orphans on either side.
  3. Close the loop - quick sanity check that nothing still references what you removed or duplicates what you consolidated.

Stop conditions

  • Do not expand into unrelated refactors or features the user didn't imply.
  • Do not ask "should I also fix X?" for obvious siblings in the same class - just fix them and mention what you included in the summary.

Communication

Briefly state what pattern you generalized (one sentence) and what else you fixed in the same pass, so the user sees they won't need to repeat themselves.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 32 lines · 340 tokens per session scan A 9c11b4948aca

Subscribe to this mod's changes

generalize-once is a cursor rule published in the GitHub repository LiuYihey/Agent-Engineering (5 stars, last pushed 28d ago), licensed MIT. It adds 340 tokens to every session, about $0.0017 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.