Author and review GitHub Actions workflow YAML safely so syntactically-valid YAML can't ship a workflow that GitHub Actions refuses to run. USE FOR: editing, adding, or reviewing any file under .github/workflows/, writing run-name/name/if/env/run values that contain ${{ }} expressions, diagnosing a run that fails with 'This run likely failed because of a workflow file issue' and no jobs starting, deciding when a workflow scalar must be quoted, validating workflows with actionlint. DO NOT USE FOR: authoring application YAML unrelated to GitHub Actions, Azure Pipelines, GitLab CI, or non-workflow YAML. SCOPE: this skill covers *syntactic/structural* correctness of workflow YAML (quoting, parsing, actionlint); for *semantic and functional* workflow design (what a workflow should do, agentic-workflow behavior), see .github/agents/agentic-workflows.agent.md — the two are complementary. INVOKES: actionlint (downloaded pinned binary) plus git/grep for inspection.
npx skills add https://github.com/dotnet/skills --skill authoring-github-workflows
GitHub Actions workflow files are YAML, but valid YAML is not the same as a valid workflow. A workflow can parse cleanly with yaml.safe_load (or a casual review) yet still be rejected by GitHub Actions at load time — producing the opaque failure *"This run likely failed because of a workflow file issue"* with zero jobs started. This skill teaches the YAML-vs-Actions traps (the #-as-comment trap above all), how to quote expression scalars correctly, and how to validate with actionlint before merge.
> Scope: syntactic vs. semantic. This skill is about the *syntactic and structural* correctness of workflow YAML — quoting, parsing, and actionlint-level validity that determines whether GitHub Actions will load and run a file at all. It is not about *what* a workflow should do or how an agentic workflow should behave. For *semantic and functional* guidance (designing workflow logic, agentic-workflow patterns, gh-aw authoring), use .github/agents/agentic-workflows.agent.md. The two are complementary: get the behavior right with the agent, get the YAML right with this skill.
.github/workflows/.run-name, name, if, env, with, or run value that embeds a ${{ }} expression.main suddenly breaks for every run after a workflow edit merged, even though the change "looked fine."run: block (that is a scripting task, not a workflow-syntax task).# inside an unquoted expression becomes a YAML commentIn YAML, a space followed by # starts a comment. In an unquoted (plain) scalar, everything from that space-then-# to end-of-line is silently discarded:
# BAD — the run-name is silently truncated at " #"
run-name: ${{ inputs.pr_number != '' && format('Evaluate PR #{0} @ {1}', inputs.pr_number, inputs.head_sha) || '' }}
YAML parses this as run-name: ${{ inputs.pr_number != '' && format('Evaluate PR — an unterminated ${{ expression. yaml.safe_load succeeds (it just sees a truncated string with a trailing comment), so the bug passes naive validation, but GitHub Actions rejects the malformed expression and refuses to start any run.
# GOOD — wrap the whole value in double quotes so '#' stays inside the scalar
run-name: "${{ inputs.pr_number != '' && format('Evaluate PR #{0} @ {1}', inputs.pr_number, inputs.head_sha) || '' }}"
The inner expression already uses single quotes, so double-quoting the scalar is safe. This is exactly the bug that broke dotnet/skills evaluation on main (PR #746 → fixed by quoting).
| Character / pattern | Why it breaks | Fix |
|---------------------|---------------|-----|
| space then # (space-hash) | Starts a YAML comment; truncates the value | Quote the whole value |
| Leading *, &, !, ?, \|, >, @, ` | YAML anchors/aliases/tags/block scalars | Quote the value |
| Leading { or [ | Parsed as flow mapping/sequence (a bare ${{ }} starts with $, which is safe, but {{ after a leading char is risky) | Quote the value |
| : then space (colon-space) inside the value | Parsed as a nested mapping key | Quote the value |
| Leading/trailing spaces that matter | Plain scalars strip them | Quote the value |
| Values that are true/false/yes/no/on/off/numbers but must stay strings | YAML type coercion | Quote the value |
Rule of thumb: if a name, run-name, if, env, or with value contains a ${{ }} expression *and* any literal #, :, or leading special character, wrap the entire scalar in double quotes.
git diff --name-only origin/main... -- .github/workflows/
For each file, scan every line that contains ${{ together with a #, a colon-space, or a leading special character.
Wrap the full value in double quotes when the value embeds an expression and contains a # or other special character (see the table above). Prefer double quotes when the inner expression uses single quotes, and vice-versa. Do not escape the ${{ }} braces — quoting the scalar is enough.
actionlint understands the GitHub Actions schema *and* the expression grammar, so it catches exactly this class of bug that plain YAML linters miss. Download a pinned release and run it:
ACTIONLINT_VERSION=1.7.7
ACTIONLINT_SHA256=023070a287cd8cccd71515fedc843f1985bf96c436b7effaecce67290e7e0757
curl -fsSLo actionlint.tar.gz \
"https://github.com/rhysd/actionlint/releases/download/v${ACTIONLINT_VERSION}/actionlint_${ACTIONLINT_VERSION}_linux_amd64.tar.gz"
# Verify the download against the pinned checksum before extracting/executing it:
echo "${ACTIONLINT_SHA256} actionlint.tar.gz" | sha256sum -c -
tar -xzf actionlint.tar.gz actionlint
# Focus on workflow/expression correctness; silence shell/py style noise:
./actionlint -shellcheck= -pyflakes= -color .github/workflows/*.yml
On Windows PowerShell, use the actionlint_<ver>_windows_amd64.zip asset and Expand-Archive.
The truncated-expression bug surfaces as:
got unexpected EOF while lexing end of string literal, expecting ''' [expression]
A clean exit code 0 means the workflows are structurally valid.
Do not rely on yaml.safe_load, yamllint, or "it parses" as proof. They accept the truncated-comment form. Only actionlint (or pushing and watching GitHub Actions parse it) validates the Actions layer.
This repository runs actionlint automatically (see .github/workflows/actionlint.yml) on any PR that touches .github/workflows/. Ensure your change passes that check before requesting review. If you add a new workflow, the gate covers it automatically.
${{ }} value containing #, a colon-space, or a leading special character is wrapped in quotes.actionlint -shellcheck= -pyflakes= .github/workflows/*.yml exits 0.actionlint CI check is green on the PR.| Pitfall | Solution |
|---------|----------|
| Unquoted run-name/name with # inside the expression | Wrap the whole value in double quotes |
| Trusting yaml.safe_load/yamllint/a code review to catch it | Run actionlint; YAML-only checks accept the truncated form |
| Escaping ${{ braces to "fix" it | Don't — quote the scalar instead; escaping breaks the expression |
| Using single quotes around a value that contains single quotes | Use double quotes for the outer scalar |
| Adding actionlint with shellcheck enabled and drowning in pre-existing shell-style warnings | Run with -shellcheck= -pyflakes= to focus on workflow/expression errors |
| Assuming a green YAML lint means the workflow will run | Push and confirm jobs actually start, or rely on the actionlint gate |
.agents/skills/create-skill/SKILL.mdReviews and authors Cloudflare Workers code against production best practices. Load when writing new Workers, reviewing Worker code, configuring wrangler.jsonc, or checking for common Workers anti-patterns (streaming, floating promises, global state, secrets, bindings, observability). Biases towards retrieval from Cloudflare docs over pre-trained knowledge.
Azure AI Text Translation SDK for real-time text translation, transliteration, language detection, and dictionary lookup. Use for translating text content in applications.
Use when editing worker/src/constants/default-model-prices.json, packages/shared/src/server/llm/types.ts, pricing tiers, tokenizer IDs, or matchPattern regexes for OpenAI, Anthropic, Bedrock, Vertex, Azure, or Gemini model pricing.
Use when contributing to the Cloudflare Docs repository — writing or editing documentation pages, choosing content types or components, adding changelog entries, reviewing docs, or learning how to contribute.
网文写作工具集基础设施部署。为 Claude Code / OpenCode / Codex / ZCode / OpenClaw / Reasonix 提供内置适配;Web AI / 通用 Agent 可走 skills + AGENTS.md 文件模式。触发方式:/story-setup、$story-setup、「准备写书」「帮我搭一下环境」「配置写作项目」。
| Azure AI Text Translation SDK for real-time text translation, transliteration, language detection, and dictionary lookup. Use for translating text content in applications.
Build translation applications using Azure Translation SDKs for JavaScript (@azure-rest/ai-translation-text, @azure-rest/ai-translation-document). Use when implementing text translation, transliteration, language detection, or batch document translation.
KQL language expertise for writing correct, efficient Kusto Query Language queries. Covers syntax gotchas, join patterns, dynamic types, datetime pitfalls, regex patterns, serialization, memory management, result-size discipline, and advanced functions (geo, vector, graph). USE THIS SKILL whenever writing, debugging, or reviewing KQL queries — even simple ones — because the gotchas section prevents the most common errors that waste tool calls and cause expensive retry cascades. Trigger on: KQL, Kusto, ADX, Azure Data Explorer, Fabric Real-Time Intelligence, EventHouse, Log Analytics, log analysis, data exploration, time series, anomaly detection, summarize, where clause, join, extend, project, let statement, parse operator, extract function, any mention of pipe-forward query syntax.
Take dotnet/authoring-github-workflows from the repository into ~/.claude/skills for personal
use, or into .claude/skills inside a project.
The agent identifies a skill by the name field in its header. Two skills with the
same name cannot sit side by side — one of them will be ignored.