pull_request

A command that opens a draft pull request, a request to merge code changes, and guides a standards review with pauses for confirmation.

In plain words
What is it for?
Use it to check the branch and target, open or update a draft pull request, record manual testing, and optionally skip the walkthrough or mark the request ready.
Why use it?
It creates a documented review flow with testing notes and checkpoints before the changes are marked ready.

Command for Claude Code

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 commands/osiris-devworks/smart-citizen/pull_request
Clone the repo
git clone --depth 1 https://github.com/Osiris-DevWorks/smart-citizen

Made for: Claude Code.

Per session 25 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,981 The whole file, excluding the scripts and references it only reads on demand.
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 $0.00025 $0.01981
Opus 5 $0.00013 $0.00991
Sonnet 5 $0.00005 $0.00396
Haiku 4.5 $0.00003 $0.00198

Measured 2d ago against content hash daf5ee42db3c, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

pull_request 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.

.claude/commands/pull_request.md · 138 lines

How it starts

The opening of the file, as written. The whole thing — 138 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/pull_request

Open or update a draft PR for the current branch, run Smart Citizen's documented self-review with checkpoints between each section, then offer to mark it ready for review.

Work through every section in order. After each section, stop at the CHECKPOINT and wait for the user to confirm before continuing. Never auto-fix a standards violation without confirming first.

Arguments

$ARGUMENTS may contain:

  • --testing "description" — manual-test notes the user has already performed. These flow into the Testing performed block of the PR body.
  • --skip-draft — open the PR as ready-for-review immediately and skip the self-review walkthrough. Use only when the user has run the checks elsewhere or is intentionally bypassing the gate.

If no flags are passed, proceed with the default draft-then-walkthrough flow.

1. Branching model

  • Confirm the current branch is not main. PRs target the active release/X.Y.Z integration branch, never main directly.
  • Identify the active integration branch: the highest release/X.Y.Z that exists on origin (git ls-remote --heads origin 'release/*'). The PR's base must match.
  • The version suffix signals scope. A patch branch (e.g. release/1.4.2) is reserved for bug fixes and minor polish. If the diff includes new features or a meaningful behavior change, flag the scope mismatch — do not proceed silently.

CHECKPOINT — confirm the branching model setup before proceeding. Wait for the user.

2. PR scope

A PR should cover one concern: a single feature, a single bug fix, a batch of related bug fixes (same subsystem, same release-branch scope), a self-contained refactor, or a self-contained doc update. Mixed-concern PRs create review burden — surface candidate splits and let the user decide.

Run these against git diff --name-only $(git merge-base HEAD origin/main)...HEAD plus git diff --shortstat:

  • File-count signal: more than 8 files touched outside a deliberate cross-cutting change (project-wide rename, doc reorganization, dep bump) → flag for scope review.
  • Net-line signal: more than ~400 net lines changed, excluding generated files and lock files → flag for scope review.
  • Concern-mixing signal: scan paths and patches. If any of these are mixed in one diff, list each cluster with its file list and ask the user to confirm or split:
    • feature work in one subsystem + unrelated bug fix in another
    • refactor of existing code + a new feature built on top of it (refactor first, ship, then build)
    • doc rewrite unrelated to the code change
    • "drive-by" formatting or rename edits in files not otherwise needed for the work
  • Title sanity: if the working title needs "and" or a semicolon to describe both concerns, that's a split signal.

Read the full file on GitHub · 138 lines

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 · 138 lines · 25 tokens per session scan A daf5ee42db3c

Subscribe to this mod's changes

pull_request is a command published in the GitHub repository Osiris-DevWorks/smart-citizen (99 stars, last pushed 3d ago), licensed Apache-2.0. It adds 25 tokens to every session and 1,981 once invoked, about $0.0001 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 2026-08-30.