- Home
- Skills
- Technical Documentation
- Markdown Release Guard
Markdown Release Guard
Batch-check Markdown code fences, preview safe repairs, and gate docs in CI. Local Node.js runner; no API keys.
$9
Works with the AI Agents you already use
Markdown Release Guard
Example session with this skill installed
Use Markdown Release Guard to check the bundled examples/unclosed.md and preview a safe repair without changing the file.
- Read your context and instructions
- Compiled the markdown release guard
Local demo: one broken fence detected
I checked the bundled examples/unclosed.md with the packaged Node.js runner. The initial check exited with code 1 and reported:
| File | Line | Column | Severity | Finding |
|---|---|---|---|---|
| examples/unclosed.md | 3 | 1 | error | UNCLOSED_FENCE |
The code block opened on line 3 is not closed before the end of the input. It is the only diagnostic, so the conservative repair is eligible.
Proposed replacement
# Incomplete AI response
```js
console.log('This block needs a closing fence');
```
The preview appended one closing triple-backtick fence. Revalidation found zero warnings: two fences, one code block, and zero unclosed fences. Preview mode exited with code 0. filesChanged is 1 and filesWritten is 0: the original file remains unchanged.
Next step
If applying this repair is part of the documentation task, run node scripts/guard.mjs fix examples/unclosed.md --write, inspect the diff, and recheck with --strict. This result verifies code-fence structure only; it does not establish that the example code is correct. This demonstration was reproduced locally from the shipped ZIP, not through Agensi hosted delivery.
Connects securely to your tools. The creator never sees your data.
About this skill
A missing closing fence can turn the rest of a README into a code block. Give your agent a repeatable local check before handing off documentation.
Included
- SKILL.md workflow and standalone Node.js batch runner
- File-and-line JSON diagnostics for .md, .mdx, and .markdown
- Non-writing repair previews and explicit --write mode
- Conservative repairs, revalidation, and CRLF preservation
- GitHub Actions template, demo inputs, and quickstart
- Bundled LLM Output Guard engine and MIT license
Workflow
Check files or folders, review diagnostics, preview repairs, then apply changes within your task. Strict mode fails on any warning. Ambiguous nested fences and longer fences stay for manual review.
Requirements and limits
Requires Node.js 20+ and a client that can execute local scripts with access to the complete package. No npm installation, API keys, or network calls are needed by the runner. Local ZIP execution is tested; Agensi hosted delivery of executable supporting files is not yet verified.
This is a heuristic backtick-fence checker, not a full Markdown or MDX parser. Tilde fences and container syntax are not comprehensively supported. It does not validate facts, links, or code behavior. An appended closer cannot recover the intended boundary of swallowed prose.
MIT licensed, including the engine and workflow. Purchase provides a ready-to-use package; no exclusivity or indefinite support entitlement.
Heuristic backtick-fence checker, not a complete CommonMark or MDX parser. Tilde fences and container syntax are not comprehensively supported. Automatic repair requires exactly one unclosed triple-backtick block with no other diagnostics; longer fences, mixed line endings, and ambiguous nesting require review. Does not verify facts, links, code correctness, or intended prose boundaries. Local ZIP execution tested; hosted executable supporting-file delivery remains unverified.
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 any payment.
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 1 month ago
Requires
Permissions
File Scopes
Reads user-selected Markdown files and directories. Writes only in fix --write mode. Preview is the default. Runner makes no network requests and does not execute document snippets. GitHub Actions is optional and downloads official actions/Node during CI setup.