opening-prs/SKILL.md
Open a GitHub PR via API as a flowing graph — branch + push + create_pr + mergeable poll, with structural protection against pushing to main/master/etc. Use when a Claude Code or Claude.ai container needs to land changes on GitHub via the API (no git CLI required) and the prose "create branch, push, open PR, wait for mergeable" workflow keeps drifting under context pressure.
npx skillsauth add oaustegard/claude-skills opening-prsInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
A flowing graph that turns the imperative "create branch, push files,
open a PR, poll mergeable_state" workflow into a structural DAG. The
"NEVER push directly to main" rule is encoded as a validate= gate that
physically can't be skipped.
from opening_prs import open_pr
result = open_pr(
repo="owner/repo",
branch_name="feat/cool-thing",
title="Add cool thing",
body="## Summary\n\n...",
files=[
("src/cool.py", "<file content>"),
("docs/cool.md", "<file content>"),
],
base="main", # default; protected names are also rejected
)
print(result["pr_url"]) # https://github.com/.../pull/N
print(result["mergeable_state"]) # clean | dirty | unstable | behind | blocked
The gh pr create workflow (or hand-rolled API calls) has five steps in
prose form:
mergeable_state until GitHub finishes computing itEach step has known failure modes:
main — the diagnosed pattern that
motivated github-procedures §6 in the first place.mergeable_state: null immediately after
creation. Need to poll. "Wait a few seconds and check" is prose, not a
procedure — so it gets skipped under context pressure.This skill encodes both as flowing primitives:
determine_branch ──▶ guard ──▶ get_base_head ──▶ create_branch
│ │
│ ▼
│ push_files
│ │
│ ▼
│ create_pr
│ / \
│ ▼ ▼
│ wait_mergeable present_pr [terminal]
│
└─ validate=must_not_be_base_branch
validate=must_not_be_base_branch rejects branch_name in {"main", "master", "trunk", "production", "prod"} — case-insensitive
— and the configured base itself. The body of create_branch never
fires for a protected name. No GitHub API call happens for an
invalid branch name.
retry_until=lambda r: r["mergeable_state"] in SETTLED_STATES
consumes the retry budget while GitHub computes the merge result.
unknown and null keep polling; settled states (clean, dirty,
unstable, behind, blocked) stop. Exhaustion is soft: the PR is
still presented with the last observed state.
Requires GH_TOKEN (or GITHUB_TOKEN) in the environment. Classic PAT
or fine-grained PAT with repo scope.
import os
os.environ["GH_TOKEN"] = "ghp_..." # or load from a .env file
The skill sends User-Agent: opening-prs on every API call. (GitHub
returns 401 "Bad credentials" without a UA, regardless of token
validity. Common trap.)
{
"pr_url": "https://github.com/owner/repo/pull/N",
"pr_number": 42,
"pr_state": "open",
"branch": "feat/cool-thing",
"base": "main",
"head_sha": "abc123...",
"mergeable_state": "clean",
"files_pushed": ["src/cool.py", "docs/cool.md"],
"detached_failures": [],
}
Raises RuntimeError only if the main DAG fails (validate, branch
creation, file push, or PR creation). Mergeable polling exhaustion is a
soft failure — the PR exists, the field just isn't computed yet.
open_pr(
...,
mergeable_poll_retries=8, # default 8
mergeable_poll_base_ms=2000, # default 2s
mergeable_poll_max_ms=8000, # default 8s (cap on exponential backoff)
)
git push if you
have the CLI.flowing — the DAG runner this skill is built onclosing-issues — the symmetric "close + synthesize" flowaccessing-github-repos skill for byte-layer GitHub access patternsdevelopment
--- name: verifying-claims description: Check that a document's claims about code are actually true by reading the prose, the code, and the tests and reporting (or fixing) where they disagree. Use whenever the user wants to verify a README, guide, spec, or docstring still matches the code; whenever they mention documentation drift, doc-code sync, "is this still accurate", stale docs, or keeping docs/tests/code consistent; before publishing or merging a docs change; or as a periodic doc-accuracy
tools
Query, filter, and transform Markdown structurally with mq — a jq-like CLI for Markdown. Use to extract headings/sections/code-blocks/links from .md files, build a table of contents, pull code blocks of a given language, slice or reshape LLM prompt/output Markdown, or batch-transform docs. Triggers on "extract sections from this markdown", "get all the code blocks", "jq for markdown", "mq", or any structural query over Markdown that grep/Read can't do cleanly.
development
Composes single-file HTML artifacts (PR review writeups, status reports, incident postmortems, slide decks, design systems, prototypes, flowcharts, module maps, feature explainers, kanban boards, prompt tuners) from a small JSON spec instead of hand-written HTML/CSS/JS. Use when the user asks to "compare options side-by-side", requests an HTML version of a report or review or deck, asks for a flowchart, status update, postmortem, design system reference, interactive prototype, custom editor — or explicitly says "HTML artifact", "single HTML file", "self-contained HTML". Skip for ad-hoc HTML snippets (forms, emails, embedded widgets) where there's no template fit.
development
DAG workflow runner that encodes control flow in code, not prose. Use when a procedure has 3+ steps with branching, retries, or validation that must be enforced — gates as `when=`, edge contracts as `validate=`, predicate loops as `retry_until=`. The runner owns the graph; the LLM provides leaves. Also covers parallel execution, checkpoint resume, detached side-effects.