Markdown Scanner

markdown-scanner

Internal sub-agent. This agent is not user-invokable. It is orchestrated automatically by markdown-a11y-assistant during parallel scan phases. You do not need to invoke it directly.

What It Does

markdown-scanner receives a single Markdown file path and scans it across all 9 accessibility domains, returning a structured findings object with per-domain issue lists, severity classifications, and a computed score.

When It Runs

The markdown-a11y-assistant dispatches one instance of this agent per discovered file during Phase 2 of the audit. All instances run simultaneously.

Domains Scanned

Each domains scanned, with its #, domain, WCAG and finds.

#	Domain	WCAG	Finds
1	Links	2.4.4	Ambiguous text: "click here", "read more", "here", "this", bare URLs
2	Alt Text	1.1.1	Missing alt on ![](), file-name alt text, uninformative alt text
3	Headings	1.3.1	Skipped levels, multiple H1s, duplicate heading text
4	Tables	1.3.1	No preceding description paragraph, complex tables without summaries
5	Emoji	1.3.3	Emoji in headings, emoji as sole bullet content, consecutive emoji sequences
6	Diagrams	1.1.1	Mermaid blocks without preceding text description, ASCII art without description
7	Em-dashes	Cognitive	— and – without spaces (screen reader concatenation)
8	Anchors	2.4.4	[text](#anchor) links where the target heading does not exist in the file
9	Plain Language	Cognitive	Passive voice density, unexpanded acronyms on first use, very long sentences

Output Format

The agent returns a structured findings block the orchestrator aggregates:

### Scan: path/to/file.md

**Score:** 82 (B)

| Domain | Severity | Line | Issue | Suggested Fix |
|--------|----------|------|-------|---------------|
| Links | Serious | 14 | "click here" - non-descriptive link text | "Download the configuration guide" |
| Headings | Critical | 32 | H2 followed by H4 - skipped H3 | Add an H3 section or change H4 to H3 |
| Emoji | Moderate | 5 | an emoji (for example a rocket) in an H1 heading - breaks anchor IDs | Remove emoji or replace with text |

Scoring Formula

File Score = 100 - (Critical x 15 + Serious x 7 + Moderate x 3 + Minor x 1)
Floor: 0