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.
npx agentmods add rules/prisma/prisma-next/optimize-for-human-time-on-prsgit clone --depth 1 https://github.com/prisma/prisma-nextWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/rules/prisma/prisma-next/optimize-for-human-time-on-prs)<a href="https://agentmods.dev/rules/prisma/prisma-next/optimize-for-human-time-on-prs"><img src="https://agentmods.dev/badge/rules/prisma/prisma-next/optimize-for-human-time-on-prs.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00685 | $0.00685 |
| Opus 5 | $0.00342 | $0.00342 |
| Sonnet 5 | $0.00137 | $0.00137 |
| Haiku 4.5 | $0.00068 | $0.00068 |
Grade A, and why
optimize-for-human-time-on-prs 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 4d 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.
This is a copy
100% identical to optimize-for-human-time-on-prs — 0 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 42 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Optimize for human time when opening PRs
The golden rule: PRs have enormous human-time overhead. Every PR costs the reviewer context-switch time, the merge-coordinator time, CI/wait-time on the operator's queue, and Linear/notification noise. A one-line PR is not "small" — its review burden has the same fixed overhead as a large PR.
Default to fewer, larger PRs, not more, smaller ones.
Concrete defaults
- Never open a one-line / one-commit / "chore" PR on its own. If the change can ride along on an in-flight PR's branch, push it there. The reviewer is already holding the context; the extra commit costs them nothing additional to look at. A separate PR costs them a full second review-cycle from scratch.
- Never split an existing branch into multiple PRs because the commits "feel separable." That value judgment belongs to the operator, who is the one paying the review cost. Same-branch / same-PR is the default; split only on explicit operator instruction.
- Ancillary work that surfaces during a slice (skill amendments, rule additions, follow-up doc tweaks, small fixes) goes on the same branch as the slice it surfaced from. Push to the existing PR. Do not create a new branch, do not rebase the ancillary commit onto
main, do not open a new PR. - If there is no in-flight PR to ride along on, and the change is genuinely a one-liner with no companion work, batch it. Stash the change locally, work on it next to whatever non-trivial work comes next, and land them together. Don't open a PR just to land the one-liner alone.
Override triggers (explicit only)
Open a new / separate PR only when the operator has explicitly said so. Examples that count as explicit:
- "open a PR for X"
- "this should be its own PR"
- "land X separately from Y"
- "don't include the skill change in the TML-XXXX PR"
These do not count as triggers:
- The change "feels like a separable concern" to the agent.
- The change touches "a different surface area."
- The change "would be cleaner on its own."
- The agent thinks the operator "would probably want it split" without having been told so.
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.
- 4d ago First seen · 42 lines · 685 tokens per session scan A 1cb1d1b1b571
optimize-for-human-time-on-prs is a cursor rule published in the GitHub repository prisma/prisma-next (419 stars, last pushed 10d ago), licensed Apache-2.0. It adds 685 tokens to every session, about $0.0034 per session on Opus 5. A static security scan graded it A with 0 findings. It is 100% identical to optimize-for-human-time-on-prs, differing in 0 lines, and is treated as a copy.
Other cursor rules, from other repositories
cli-error-handling
CLI command error handling patterns.
prefer-direct-imports-over-module-mocks
Prefer extracting a testable core over vi.mock / vi.resetModules when unit tests need to reach production logic entangled with config, env, or singletons.
control-plane-descriptors
Control plane descriptor and instance implementation patterns.
family-instance-domain-actions
Family instance domain action implementation patterns.
import-validation
Import validation and package layering rules.
no-target-branches
Never branch on target - use adapters instead.