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 skills add jamestexas/agents --skill pr-bucketsgit clone --depth 1 https://github.com/jamestexas/agentsWrote 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/skills/jamestexas/agents/pr-buckets)<a href="https://agentmods.dev/skills/jamestexas/agents/pr-buckets"><img src="https://agentmods.dev/badge/skills/jamestexas/agents/pr-buckets/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/jamestexas/agents/pr-buckets"><img src="https://agentmods.dev/badge/skills/jamestexas/agents/pr-buckets.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00102 | $0.01404 |
| Opus 5.5 | $0.00041 | $0.00562 |
| Sonnet 5.5 | $0.00020 | $0.00281 |
| Haiku 4.5 | $0.00010 | $0.00140 |
Grade A, and why
pr-buckets 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 1mo 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.
How it starts
The opening of the file, as written. The whole thing — 102 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/pr-buckets — slice one branch into small, reviewable PR buckets
You have one dev branch you reasoned about linearly. Reviewers want small PRs.
This skill decides the bucket boundaries — the logical PR-sized slices — so
jj-stack can mechanically produce them. It decides what the PRs are; it does
not touch git/jj.
The bar
A bucket is one concern a reviewer can hold in their head in one sitting. Default threshold: ~150 LoC of implementation (see the counting rule — tests don't count the same). Treat it as a strong default, not a hard gate:
- Under threshold, single concern → ship as its own bucket.
- Over threshold because of cohesive impl (a parser + its dispatcher) → try to split by sub-concern; if it genuinely can't split without leaving a half-built thing, keep it whole and say so in the bucket note ("187 LoC — the dispatcher is atomic; splitting would ship an unwired handler").
- Over threshold because of tests → not a violation. Tests inflate line count but are cheap to review. Count impl and test separately (below).
Counting rule (impl vs test)
Report each bucket as impl / test LoC, not one number. The threshold applies
to impl. A 357-line bucket that is 133 impl + 224 table-test is a small
PR wearing a big number — flag it, don't split it. Get the split with:
git show --stat <commit> # per-file line counts
jj diff -r <rev> --stat
Sum the non-_test.go / non-*_test.* files for impl; the rest is test.
Where the boundaries are (in priority order)
- By layer / module — a change that spans
core/+server/+cmd/is usually 2–3 buckets (the dispatcher, the wiring, the entrypoint), not one. - By concern — a new capability + a test-harness change that enables testing it are two buckets (the feature, and the seam). "My test is separate" is a first-class boundary.
- By file disjointness — buckets whose file sets don't overlap split
cleanly in jj (
jj split <paths> -m "<msg>"is fully non-interactive when slices are file-disjoint). A bucket boundary that cuts within a file forces hunk-level splitting, which is a different and slower tool. Prefer boundaries that fall on file lines when you can. - By blast radius — a shared-signature change (a constructor that gains an arg) and its call-site updates must land together or the intermediate commit won't build; that forces a bucket boundary (they're one bucket) or a back-compat wrapper (they're two).
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.
- 1mo ago Changed · +9 lines 71b8fd598194
- 1mo ago First seen · 93 lines · 102 tokens per session scan A c67824dcead7
pr-buckets is a skill published in the GitHub repository jamestexas/agents (2 stars, last pushed 23d ago), licensed Apache-2.0. It adds 102 tokens to every session and 1,404 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-09-11.
Other skills, from other repositories
commit-messages
Write git commit messages and pull request titles in the WARP format used by the swarms repo.
setup-pre-commit
A setup guide for Git pre-commit checks, which run automatically before a commit is created. It configures Husky, lint-staged, Prettier, type checking, and tests in a repository.
re0-git
Rewrite a finished commit's message into a clean, handoff-ready form in your own log style, so git log alone tells the story. User-invoked: run it after a commit.
commit-and-push
Create a well-formatted git commit and push to remote repository.
customize-contribute-back
Contribute a user's AIWG customization back upstream as a PR — reviews for general applicability, creates branch, opens PR.
effect-ledger
Record an effect, sign an effect, reconcile an effect or check "did this PR merge" through the signed AIWG effect ledger, using one aiwg effect command with scriptable exit codes.