More screenshots
Works with the AI Agents you already use
Explain This PR
Understand any pull request fast: a short brief first, then a walkthrough of behavior, intent and fit, with fact and inference kept apart. Prepares you to review a teammate's PR or to explain your own.
$5
Explain This PR
Example session with this skill installed
I'm about to review https://github.com/rikuws/mood./pull/8. Walk me through what it does before I start. Read-only: don't post anything.
- Read your context and instructions
- Compiled the explain this pr
Mode: Reviewer prep. PR #8 in rikuws/mood. at head f5817f3 ("Stack 1 of 4", merge order #8 → #9 → #10 → #11).
Purpose - Before this PR a saved item had one text field holding both captured source text (an X post body, a page description) and whatever the user typed as a "Note". This PR adds a separate user note and a keep / reject verdict, so an agent can tell the user's reasoning from the source.
Parts - 1. Model … 2. Persistence and sync … 3. Agent API and docs … 4. Edit sheets on Mac and iOS …
The key question - What happens to note / verdict when something other than this PR's editor rewrites an item: capture sheets, sync dedupe, an older build?
Start reading at - Shared/Models/Inspiration.swift, init(from:).
…
Worth remembering: a concrete trace. The user opens a saved X post, types "Love the restrained grid." under "Your note" and picks Keep → updateInspiration trims the note and stores verdict = .keep → mood-agent inspiration --id <uuid> returns "note": "Love the restrained grid.", "verdict": "keep", with the post body still in text.
…
Where to look hard
Sync/Sources/SyncMerger.swift, fillMissingMetadata - nil is both "unset" and "cleared / Undecided", so a duplicate merge can copy an old verdict or note back over one the user just cleared.
- Old builds and
updatedAt- …
Documentation/AGENT_API.md - the "don't conflate them" contract versus capture sheets that still put the user's words in text.
adversarial-pr-review is available: want me to run it on these spots?
Connects securely to your tools. The creator never sees your data.
About this skill
Built by a dev who reviews PRs daily. Most review time goes into working out what a change is for. Explain This PR does that with you: the big picture first, then one piece at a time, so you can describe the change in your own words before you judge it.
What you get
- A brief of about 200 words: the purpose, the parts, the one behaviour change, the key question, and where to start reading.
- A walkthrough in a sensible order: foundations, then logic, then wiring, then UI. Refactors are kept apart from behaviour changes, and each piece covers what it does, how it fits the codebase, and what's worth remembering.
- Fact kept apart from guesswork. What the code shows, what it infers about intent, and what it couldn't verify are always labelled.
Modes
- Reviewer prep is the default for a PR. It ends with "Where to look hard": the 2-3 riskiest spots and the question to answer before approving.
- Your own PR explains your diff the way a reviewer will see it. It flags files that look accidentally touched, lists the questions a reviewer is likely to ask, and gives you a summary ready for the PR description.
- Learning is paced, one piece per message, for onboarding or catching up on what changed since last week.
Post it to the PR as a comment with collapsible sections, only if you ask.
It's read-only by default and works with GitHub PRs through gh, or with any branch, commit or diff through git. It pairs with Adversarial PR Review.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse All Claude Skills.
Explains rather than judges: it isn't a code review and only briefly mentions real problems it notices (pair it with adversarial-pr-review for that). History claims are limited to what git log, blame and linked issues show. Posting a PR comment is opt-in only. Check-yourself questions appear only in learning mode.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 22 days ago
- Accounts
GitHub
- Software
Git
GitHub CLI
- Optional
- ripgrep
Sign in to the GitHub CLI for GitHub pull requests; branches, commits and local diffs need only Git.
Permissions
Allowed Hosts
File Scopes
Read-only by default: never edits, commits, pushes or posts unless you explicitly say yes. Posting the walkthrough as a PR comment is opt-in and shown in full first.