Skip to content
Skills Library
Programming · Automation

pr

/pr

DraftForge PR lifecycle workflow. Use when the user runs /pr, opens or pushes a pull request, or asks to monitor CI checks and Copilot review on a branch. Creates the PR (or reuses an existing one), polls GitHub Actions and Copilot up to ~15 min, then analyses CI failures and Copilot comments and reports findings grouped by file and skill.

Repository
DraftForge ↗
Licence
GPL-3.0
Last repository change
25.10.2026

Assessment and security

100/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content35 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
4
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 25.10.2026, without references to retired models or APIs.

Licence

GPL-3.0, otwarta.

What it does and when to use it

DraftForge PR lifecycle workflow. Use when the user runs /pr, opens or pushes a pull request, or asks to monitor CI checks and Copilot review on a branch. Creates the PR (or reuses an existing one), polls GitHub Actions and Copilot up to ~15 min, then analyses CI failures and Copilot comments and reports findings grouped by file and skill.

Original author description

DraftForge PR lifecycle workflow. Use when the user runs /pr, opens or pushes a pull request, or asks to monitor CI checks and Copilot review on a branch. Creates the PR (or reuses an existing one), polls GitHub Actions and Copilot up to ~15 min, then analyses CI failures and Copilot comments and reports findings grouped by file and skill.

How it works

  1. Purpose
  2. When to trigger
  3. Workflow
  4. What this skill does NOT do
  5. References
Original SKILL.md headings · 87 lines, 708 words
SKILL.md preview
---
name: pr
description: DraftForge PR lifecycle workflow. Use when the user runs /pr, opens or pushes a pull request, or asks to monitor CI checks and Copilot review on a branch. Creates the PR (or reuses an existing one), polls GitHub Actions and Copilot up to ~15 min, then analyses CI failures and Copilot comments and reports findings grouped by file and skill.
---

# /pr — PR lifecycle workflow

## Purpose

Drive a PR from "branch pushed" to "actionable review summary" without the user babysitting it. The skill:

1. Resolves PR state (open existing or create new).
2. Requests Copilot as a reviewer if missing.
3. Polls GitHub Actions checks and the Copilot review (up to 15 min).
4. Pulls CI failure logs and Copilot inline comments.
5. Categorises findings by file → matched `.github/instructions/*.instructions.md` → canonical skill, then surfaces them to the user with concrete next steps.

## When to trigger

- User types `/pr`.
- User says "open a PR" / "create a PR" / "spin up a PR" on the current branch.
- User says "watch CI" / "monitor checks" / "wait for Copilot" on a branch with an open PR.
- A push has just happened and the user is asking what's next.

Do **not** trigger automatically just because a branch has uncommitted work — confirm first.

## Workflow

### Phase 1 — Resolve state

```bash
branch=$(git branch --show-current)
pr_json=$(gh pr view --json number,state,headRefName,baseRefName,url,reviewRequests,reviews 2>/dev/null || echo "")
```

- If `pr_json` is empty → no PR yet → go to Phase 2 (create).
- If PR exists and is `OPEN` → go to Phase 3 (request Copilot if absent, then monitor).
- If PR is `MERGED` / `CLOSED` → report and stop; ask whether to open a new one.

Before creating, verify the branch is pushed (`git rev-parse --abbrev-ref --symbolic-full-name @{upstream}` succeeds) and is ahead of `main`.

### Phase 2 — Create the PR (if missing)

Pre-fill title/body from commits when the branch has 1–3 commits; otherwise prompt the user once for a title. Always use a HEREDOC body with `## Summary` and `## Test plan` sections, per repo convention. Full patterns and the exact `gh pr create` invocation: [references/creating-the-pr.md](references/creating-the-pr.md).

### Phase 3 — Request Copilot reviewer (idempotent)

```bash
PR=$(gh pr view --json number -q .number)
# Skip if Copilot is already in reviewRequests OR has already reviewed
need_copilot=$(gh pr view $PR --json reviewRequests,reviews \
  --jq '(.reviewRequests + .reviews) | any(.login // .author.login | tostring | test("copilot"; "i")) | not')
if [ "$need_copilot" = "true" ]; then
  gh pr edit $PR --add-reviewer @copilot
fi
```

The exact bot login varies (`copilot-pull-request-reviewer[bot]` / `Copilot`). The case-insensitive `test("copilot"; "i")` match is intentional — don't hardcode an exact login.

### Phase 4 — Monitor checks and Copilot review

Poll every 30 s, max 30 iterations = 15 min. Break early when **both** are done. Full polling pattern, jq queries, and "is Copilot done?" heuristics: [references/monitoring-ci-and-copilot.md](references/monitoring-ci-and-copilot.md).

While polling, narrate progress in **single-line updates** ("iter 4/30: 3 checks running, copilot pending"). Do NOT print large JSON blobs to the user — those go to logs.

If the 15 min budget exhausts:
- If checks are still running: report which jobs and proceed to analysis with what completed.
- If Copilot is still pending: report and analyse CI-only; tell the user they can re-run `/pr` later to pick up the Copilot review.

### Phase 5 — Analyse and report

Two analyses run in parallel; one merged report:

1. **CI failures** — `gh run view <id> --log-failed` for each failed run, extract the error lines (look for `FAIL`, `Error`, `Traceback`, `assert`, lines ending with `:[0-9]+:` for compiler errors). Cap at 10 lines per failed step.
2. **Copilot inline comments** — `gh api repos/{owner}/{repo}/pulls/{N}/comments` → for each comment, match the `path` against every `.github/instructions/*.instructions.md` `applyTo` glob and tag it with the matched skill(s).

Output format and severity heuristics: [references/analyzing-results.md](references/analyzing-results.md).

## What this skill does NOT do

- Does not push for the user (assumes branch is already pushed; tells the user to push if not).
- Does not fix the issues Copilot raises — only surfaces them with the canonical-skill pointer so the user / next skill invocation can act.
- Does not enable automatic Copilot review at the repo level — that's a one-time UI step (Settings → Rules → Rulesets).
- Does not stream raw `gh run view --log-failed` output into the chat — extracts just the error lines.

## References

- [creating-the-pr.md](references/creating-the-pr.md) — `gh pr create` patterns, title/body autofill, push-check
- [monitoring-ci-and-copilot.md](references/monitoring-ci-and-copilot.md) — polling loop, jq queries, "done" heuristics
- [analyzing-results.md](references/analyzing-results.md) — CI log extraction, Copilot-comment → skill matching, output format
Files in this package (4)
  • SKILL.md 5 KB
  • references/analyzing-results.md 5 KB
  • references/creating-the-pr.md 2 KB
  • references/monitoring-ci-and-copilot.md 3 KB

Versions

RepositoryScoreLicence
This page · kettleofketchup/DraftForge100GPL-3.0
harlan-zw/harlan-claude-code100MIT
harlan-zw/harlan-agent-kit100MIT
jongio/skills100MIT
windmill-labs/windmill100Apache-2.0
opsmill/infrahub100Apache-2.0
nesquikm/dev-process-toolkit100MIT
tony/skills100MIT
tony/skills100MIT

Database snapshot: 13.11.2026 · package version 16a9c57c7df8 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100copilot-check
    Check Copilot Review Comments. Fetches, adversarially verifies, and resolves Copilot code-review comments on the current PR, then drives an automated re-review loop (re-request → watch → verify → fix → converge). Use when the user asks to "check copilot comments", "what did…
  • 92github
    GitHub CLI operations for issues, PRs, reviews, CI, and Copilot workflows. Use when working with pull requests, issues, CI status, code review, GitHub Copilot, triage, or PR shipping.
  • 100github-pr-workflow
    Use when creating, updating, or managing pull requests — automates the full PR lifecycle (open, review requests, labels, merge) via GitHub MCP
  • 100pr-review-comments
    Guide for submitting/posting inline PR review comments to GitHub. Use this when you need to post code review comments on specific lines of a Pull Request.
  • 100create-pr
    Runs the full PR workflow — creates a feature branch, commits, pushes, and opens the pull request. Use when the user asks to create or open a PR ("create a PR", "push this up and open a PR").
  • 100copilot-address-reviews
    Evaluate and address GitHub Copilot PR review items for a GitHub pull request. Use when user provides a GitHub PR link and wants Copilot review comments checked, verified, fixed, committed, and summarized. Fetch review items with `gh`, verify each claim against source/tests…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.