majiayu000 Validating Markdown Links
This Skill provides comprehensive guidance for validating markdown links across the repository, including internal link validation, external link verification, Hugo-specific patterns, and checker implementation strategies.
Purpose
Use this Skill when:
- Implementing link validation in checker agents
- Validating internal documentation links
- Verifying external URL accessibility
- Checking Hugo site link formatting
- Implementing link caching strategies
- Understanding link validation patterns
Link Validation Principles
Why Link Validation Matters
Broken links:
- Break user navigation experience
- Reduce documentation credibility
- Create maintenance burden (hard to find broken links manually)
- Indicate structural issues (moved/deleted files)
Link validation:
- Prevents broken links from reaching production
- Catches file renames and moves
- Validates external resources still exist
- Ensures consistent link formatting
Validation Scope
What to validate:
- Internal markdown links (docs/, governance/, plans/)
- Hugo content links (apps/ayokoding-web/, apps/ose-platform-web/)
- External URLs (HTTP/HTTPS)
- Image links (relative paths)
- Anchor links (same-page headings)
What NOT to validate:
- Links in code blocks (examples, not actual links)
- Links in quoted text (preserved formatting)
- Commented-out links (intentionally disabled)
Internal Link Validation
Required Link Format
Documentation files (docs/, governance/, plans/, root .md files):
✅ PASS: File Naming Convention ✅ PASS: AI Agents Convention
❌ FAIL: File Naming Convention ← Missing .md extension ❌ FAIL: [[file-naming]] ← Obsidian wiki link (not GitHub-compatible) ❌ FAIL: file-naming.md ← Using filename as link text
Hugo content files (apps/ayokoding-web/content/, apps/ose-platform-web/content/):
✅ PASS: Learn AI ✅ PASS: Chat with PDF Tutorial
❌ FAIL: Learn AI ← Should omit .md extension in Hugo ❌ FAIL: Learn AI ← Should use absolute path in Hugo (starts with /)
Validation Methodology
Step 1: Extract Links from Markdown
Use regex or markdown parser to extract all links
Step 2: Categorize Links
Separate into categories:
- Internal links (start with ./, ../, or /)
- External links (start with http://, https://)
- Anchor links (start with #)
- Image links (extensions: .png, .jpg, .svg, etc.)
Step 3: Validate Internal Links
For each internal link:
- Resolve relative path from current file location
- Check target file exists using filesystem
- Validate format (has .md extension for docs/, no extension for Hugo)
- Check link text quality (descriptive, not filename-based)
Common Internal Link Errors
Error 1: Missing .md extension
❌ FAIL: File Naming ✅ PASS: File Naming
Criticality: HIGH - Breaks GitHub web navigation Detection: Check link ends with .md (docs/ files only)
Error 2: Wrong relative path depth
From: governance/conventions/formatting/linking.md (3 levels deep) ❌ FAIL: Documentation Home ← Only 1 ../, need 3 ✅ PASS: Documentation Home
Criticality: CRITICAL - Link points to wrong file or 404 Detection: Resolve path and check file exists
Error 3: Obsidian wiki links
❌ FAIL: [[file-naming-convention]] ❌ FAIL: [[file-naming-convention|File Naming]] ✅ PASS: File Naming Convention
Criticality: HIGH - Breaks GitHub web compatibility Detection: Regex match for wiki-style links
Error 4: Filename as link text
❌ FAIL: ex-co__file-naming.md ✅ PASS: File Naming Convention
Criticality: MEDIUM - Poor accessibility and readability Detection: Check if link text matches filename pattern or contains file extension
Error 5: Hugo link with .md extension
From: apps/ayokoding-web/content/learn/ai/_index.md ❌ FAIL: Chat with PDF ← Hugo omits .md ✅ PASS: Chat with PDF
Criticality: HIGH - Hugo won't resolve the link correctly Detection: Check Hugo content files don't use .md in absolute paths
Error 6: Hugo link using relative path
From: apps/ayokoding-web/content/learn/ai/_index.md ❌ FAIL: Security ← Hugo needs absolute paths ✅ PASS: Security
Criticality: CRITICAL - Breaks navigation when page rendered in different contexts Detection: Check Hugo content uses absolute paths (start with /)
External Link Validation
Verification Strategy
Challenge: External links can be slow to verify (network requests)
Recommended approach:
- Cache results - Store validation results in cache file
- Respect cache TTL - Re-verify after 7 days (configurable)
- Batch verification - Verify multiple URLs in parallel
- Handle failures gracefully - Network errors != broken link
HTTP Request Pattern
Verification steps:
- HEAD request first - Faster than GET, checks if URL accessible
- Follow redirects - HTTP 301/302 are OK (but report for info)
- Check status codes:
- 200-299: OK
- 300-399: REDIRECT (report but don't fail)
- 400-499: BROKEN (client error, link is wrong)
- 500-599: SERVER_ERROR (temporary, re-verify later)
- Timeout: UNREACHABLE (network issue, re-verify later)
Link Caching Strategy
Cache file format (JSON):
{ "https://diataxis.fr/": { "status": "OK", "http_code": 200, "last_checked": "2026-01-25T13:30:00+07:00", "ttl": 604800 } }
Cache TTL recommendations:
- OK links: 7 days
- BROKEN links: 1 day
- REDIRECT links: 7 days
- SERVER_ERROR: 1 hour
- UNREACHABLE: 1 hour
Common External Link Errors
Error 1: Link returns 404
Criticality: HIGH - Link is dead, user gets 404 Action: Update or remove link
Error 2: Link redirects (301/302)
Criticality: LOW - Link works but could be updated to final URL Action: Consider updating to final destination (optional)
Error 3: Link times out
Criticality: MEDIUM - May be temporary network issue Action: Re-verify after TTL expires, flag if persistent
Hugo-Specific Link Validation
Hugo Link Patterns
Hugo internal links use absolute paths without .md extension:
✅ PASS: Learn AI ✅ PASS: Chat with PDF
❌ FAIL: Learn AI ← Should be absolute path ❌ FAIL: Chat with PDF ← Should omit .md
Why Hugo Uses Absolute Paths
Hugo renders the same navigation content in different contexts (sidebar, mobile menu, homepage). Relative paths would resolve incorrectly depending on page context. Absolute paths work consistently across all contexts.
Validation Differences: Hugo vs. Docs
| Aspect | Docs/ Files | Hugo Content |
|---|---|---|
| Extension | MUST include .md | MUST omit .md |
| Path Type | Relative (../, ./) | Absolute (/path) |
| Link Format | Text | Text |
| Validation | Check file exists on filesystem | Check file exists in content/ tree |
| Criticality | CRITICAL if missing .md | CRITICAL if includes .md |
Checker Implementation Patterns
Link Validation Workflow
Standard 5-step checker workflow:
Step 0: Initialize Report Step 1: Discover Files (glob for *.md) Step 2: Extract Links (parse markdown, categorize) Step 3: Validate Internal Links (file exists, format correct) Step 4: Validate External Links (HTTP request, cache results) Step 5: Finalize Report (summary, grouped by criticality)
Progressive Writing Requirement
CRITICAL: Write findings to report file immediately after discovery (don't buffer in memory)
Why: Context compaction can lose buffered findings during long validation runs
Tool Requirements for Link Checkers
Required tools:
- Read: Read markdown files to extract links
- Glob: Find all markdown files in scope
- Grep: Extract link patterns
- Bash: File existence checks, path resolution, HTTP requests
- Write: Initialize and update report file
Categorization by Criticality
CRITICAL (Must fix before publication):
- Broken internal links (404, file not found)
- Wrong Hugo link format (includes .md or uses relative path)
- Obsidian wiki links (breaks GitHub compatibility)
HIGH (Should fix before publication):
- Missing .md extension in docs/ links
- Broken external links (404, 410)
- Filename as link text (poor accessibility)
MEDIUM (Fix when convenient):
- External link redirects (works but suboptimal)
- External link timeouts (may be temporary)
- Suboptimal link text (not descriptive enough)
LOW (Optional improvements):
- Consider updating redirect to final URL
- Suggest alternative link text
Dual-Label Pattern for Link Checkers
Link checkers use both verification status AND criticality:
Finding: [BROKEN] - Internal Link to Non-Existent File
Verification: [BROKEN] - Target file does not exist Criticality: CRITICAL - Breaks user navigation
Verification labels:
- [OK] - Link is valid
- [BROKEN] - Link target doesn't exist (404)
- [REDIRECT] - External link redirects (informational)
- [FORMAT_ERROR] - Wrong format (missing .md, etc.)
Why dual labels?
- Verification describes FACTUAL STATE
- Criticality describes URGENCY/IMPORTANCE
- Complementary information for fixer decision-making
Related Conventions
Linking Standards:
- Linking Convention - Complete linking standards
- Hugo Content Convention - Hugo linking patterns
Validation Standards:
- Repository Validation Methodology - Standard validation patterns
- Criticality Levels Convention - Criticality classification
Quality Standards:
- Content Quality Principles - Link text quality requirements
Related Skills
Validation Skills:
- repo-assessing-criticality-confidence - Criticality and confidence levels
- repo-applying-maker-checker-fixer - Checker workflow patterns
- repo-generating-validation-reports - Report format and progressive writing
Domain Skills:
- apps-ayokoding-web-developing-content - Hugo-specific linking patterns
Related Agents
Link Validation Agents:
- docs-link-general-checker - Validates links in docs/, governance/, plans/
- apps-ayokoding-web-link-checker - Validates links in ayokoding-web Hugo content