| name | create-diff |
| description | Create a YugabyteDB Phorge diff for the current branch's changes. Use when the user wants to publish their changes for review. |
Create Diff
Create a Phorge diff for the current branch using arc diff --create.
Confidentiality -- read before you write anything
Treat Phorge as public. Two reasons: (1) Phorge revisions and the
Jenkins logs they trigger are visible to anyone with company access, not
just the reviewers you @-mention; (2) anything you put into the test
code, code comments, fixtures, or golden outputs on this diff lands on
the public GitHub repo verbatim when this work is later upstreamed via a
PR — there is no separate scrubbing pass at that boundary. So apply the
public-repo rules now: diff title, summary, test plan, inline comments,
commit messages, code, tests, fixtures, golden output, and filenames all
get the same treatment as a public GitHub PR.
Never put any of the following into the diff title, summary, test
plan, Phorge comments, commit messages, code comments, test code, test
fixtures, golden output, or filenames:
- Customer-identifying data — company names, account IDs,
universe/cluster names or UUIDs, support case numbers, environment
names, region/zone names tied to a customer deployment. Substitute
with
customer-1, acme-corp, or a generic description.
- PII — real names, real emails, real phone numbers, real postal
addresses, real IP addresses (public or private). In tests use only
the documentation ranges: IPv4
192.0.2.0/24 / 198.51.100.0/24 /
203.0.113.0/24 (RFC 5737), IPv6 2001:db8::/32 (RFC 3849),
hostnames example.com / example.org / example.net (RFC 2606),
names Alice/Bob/Carol, phone numbers 555-0100–555-0199.
- Unanonymized customer schema or queries — table names, column
names, SQL text, query plans, or sample rows pulled from a real
customer report. Reconstruct a synthetic minimal reproducer with
generic identifiers (
t1, users, id, value) that demonstrates
the same defect.
- Secrets / credentials — API keys, tokens, passwords, TLS
certificates, private keys, license keys, kubeconfigs, production S3
buckets, internal Slack/JIRA/Linear URLs (other than referencing them
by ticket key like
PLAT-20518), internal hostnames, vault paths.
Use placeholders like AKIAIOSFODNN7EXAMPLE in mocks.
- YugabyteDB-internal information not yet public — unreleased
roadmap, internal SLAs, embargoed security findings, internal infra
hostnames.
applies to the test code and test data you write too** — .cc, .py,
.java, .sql, golden .out files, YAML fixtures, mock responses.
If a customer's reproducer uses acme_orders with a customer_email
column populated with real addresses, rewrite it as t1/email with
user@example.com before the test goes anywhere near a diff.
When the source is a customer report: read the original from the
internal source (JIRA / support ticket / Slack) — don't paste or link
it. Reproduce locally with synthetic inputs. Land only the synthetic
reproducer. Reference the issue by internal ticket ID (PLAT-20518)
only.
If you're unsure whether a string is sensitive, don't write it down --
ask the user. See src/AGENTS.md section on Confidentiality for the
canonical list.
This rule applies at every step below: Step 1 (commit message), Steps
3-3.5 (issue body if auto-creating), Step 4 (title), and Step 6 (summary
/ test plan).
Prerequisites
This skill requires arc (Arcanist) to be installed and configured for
the YugabyteDB Phorge instance. If arc is not available, instruct the
user to set it up.
Workflow
Step 1: Commit pending changes
arc diff --create requires a clean working copy — any unstaged or
untracked changes on the branch must be committed first.
Run git status. If there are uncommitted changes:
- Stage the relevant files (avoid
git add -A/. — pick files
explicitly so secrets or stray artifacts don't get included).
- Draft a commit message based on the changes and confirm it with the
user before committing.
- Create the commit.
If the branch already has at least one commit and the working copy is
clean, skip to Step 2.
Step 2: Run linter and confirm no lint errors
Run ./build-support/lint.sh from the repo root. Report the output to
the user.
- If there are errors, stop and fix the issues before continuing.
Do not proceed to
arc diff --create.
- If there are only warnings or advice, show them to the user and
ask whether to proceed.
- If the output is clean, continue to the next step.
Step 3: Determine tracker reference
You will need a GitHub or JIRA issue describing the issue this diff is
fixing in order to create the diff. Normally there is a one-to-one
relationship between tracking issues and diffs.
First, check which area the changes touch:
- If changes are in core DB code (
src/, java/, cmake_modules/,
build-support/), remind the user that most core DB changes should
have a GitHub issue.
- If changes are in YBA/platform code (
managed/), remind the user
that most platform changes should have a JIRA ticket.
Second, scan the current conversation history for any issue reference
the user has already mentioned (i.e., GitHub issue numbers like
#31151, JIRA keys like PLAT-20518, or URLs pointing at either). If
you find a candidate, surface it to the user and ask them to confirm
it's the one to use before asking them to supply a fresh one. Only
prompt for a new reference if no candidate is found or the user rejects
the candidate.
When asking the user for a tracking reference, acceptable forms are:
- A GitHub issue number (e.g.,
#31151)
- A JIRA ticket (e.g.,
PLAT-20518)
Offer to auto-create a GitHub issue or JIRA ticket if the user
doesn't have one yet. If they accept, invoke the /create-issue
skill to file it (GitHub issue for core DB code, JIRA ticket for
managed/ platform work), then capture the issue number / JIRA key it
reports back. Since the issue tracks this diff's work, tell
/create-issue to assign it to the invoker (@me) rather than leaving
it unassigned.
Step 3.5: Gather remaining title metadata
Prompt the user for the following, one at a time:
-
Component — the component tag that will appear in the title
(e.g., DocDB, YSQL, YBA, CDC, xCluster). If the user is
unsure, suggest a component based on the files changed (e.g.,
src/yb/ → DocDB, src/postgres/ → YSQL, managed/ → YBA).
-
Title — a short description of the change. If the user doesn't
provide one, propose a title, possibly based on the commit messages
on the branch if they are informative, and confirm it with the user.
Step 4: Build the title
Construct the title in the format:
[<#GH issue> or <JIRA>] <Component>: <Title>
Examples:
[#31151] DocDB: Fix SamplingProfilerTest.EstimatedBytesAndCount in release
[PLAT-20518] YBA: Fix full move edit universe to populate userIntentOverrides during configure call
Step 5: Determine subscribers
Based on the files changed on the branch, decide the default subscriber:
- If changes are in DB code (
src/, java/, cmake_modules/,
build-support/, etc.) → add ybase as a subscriber.
- If changes are in YBA / platform code (
managed/) → add yugaware as
a subscriber.
- If changes span both, add both
ybase and yugaware.
Confirm the subscriber list with the user before creating the diff.
Step 5.5: Cut the prose the branch added
Re-read the text the branch adds — comments, architecture/ docs,
agent docs — and apply AGENTS.md Prose
discipline.
It is a gate here because it is easy to hold at the start of a task and
gone by the end of one, and this is the last point where cutting is
free.
git diff <base>...HEAD -- '*.md' # doc prose; also skim added comments in the code diff
Commit any resulting edits on the branch before Step 6, and re-run the
linter (Step 2).
Step 6: Create the diff with arc diff --create
Confidentiality — final scrub before publishing. Phorge is treated
as public (company-wide visible + test code lands on the public GitHub
repo verbatim when upstreamed). Before composing the message file,
re-read the summary, test plan, commit messages, and the test code
being added, and confirm none of the following appear: customer names
or identifiers (universe UUIDs, account IDs, support cases, environment
names, region/zone names); PII (real names / emails / phone numbers /
postal addresses / IP addresses — use RFC 5737/3849 documentation ranges
in tests); unanonymized customer schemas (table / column / query text /
query plans / sample rows from a real customer — reconstruct a synthetic
reproducer); credentials, tokens, certificates, private keys, or license
keys; internal-only hostnames, URLs, Grafana/Slack/Linear links, or
vault paths; unreleased internal information (roadmap, SLAs, embargoed
security findings, internal infra hostnames). See the top of this skill
and src/AGENTS.md Confidentiality for the full rule. If unsure
whether a string is sensitive, don't write it down — ask the user.
Run arc diff --create without using the Phorge MCP server. Do
not amend the user's existing commit to change its title — always
pass the Phorge revision title and body via --message-file so the
local commit stays untouched.
A typical invocation:
arc diff --create --message-file <path> --reviewers <reviewers> --cc <subscribers> <base>
Pre-fill the message file with:
- Title: the constructed title from step 4
- Summary: derived from the branch commits. Say why, always —
a reviewer who has to reverse-engineer the motivation is the
expensive case — then what changed, and whatever the reader must act
on (new gflags, upgrade/rollback consequences, migration steps). Not
a narration of the diff. See
AGENTS.md Prose
discipline.
- Test Plan: ask the user for one if not obvious from the branch.
Keep it to what was actually run — it is not a place for prose.
- Reviewers: (as provided)
- Subscribers:
ybase and/or yugaware per Step 4
If tests were added or considered important to verify the change works,
add the commands for running those tests to the test plan.
Show the user the resulting diff URL (e.g., https://phorge.dev.yugabyte.com/D<id>).
Step 7: Post a "trigger jenkins" comment
After the diff is created, post a comment on the new revision with the
exact text:
trigger Jenkins
Use arc to post the comment (do not use the Phorge MCP server). For
example:
echo "trigger jenkins" | arc call-conduit differential.createcomment -- '{"revision_id": "<id>", "message": "trigger jenkins"}'
Or use whatever arc subcommand is appropriate for the local setup.
Confirm the comment was posted successfully.
Step 8: Report back to the user
Output:
- The diff URL
- The constructed title
- The subscribers that were added
- Confirmation that
trigger Jenkins was posted
Notes
- Never use the Phorge MCP server for creating the diff — use
arc CLI
only.
- Always run
./build-support/lint.sh first; a failing lint should
block diff creation.
- The title format is strict:
[<issue>] <Component>: <Title>. Don't
deviate.
ybase for DB, yugaware for YBA/platform — add both when changes
span both.
- The
trigger jenkins comment kicks off continuous integration
testing; forgetting it is a common mistake.