mention.allowed should handle @

[pelikhan]

The bot-detection workflow had an incorrect format for the mentions.allowed configuration field. The field was configured with the @ prefix:

mentions:
allowed: ["@pelikhan"] # INCORRECT
According to the specification and test examples in actions/setup/js/sanitize_content.test.cjs, the allowed list should contain usernames without the @ prefix:

mentions:
allowed: ["pelikhan"] # CORRECT

FIX the mention.allowed feature to normalize the name and remove the '@' character at compile time.

[pelikhan]

/cloclo

[github-actions]

✨ Claude Response via /cloclo

Summary

Magnifique! I've implemented the fix for the mentions.allowed configuration to normalize usernames by automatically stripping the '@' prefix at compile time. The solution is elegant, user-friendly, and backward compatible! 🎭

What I Did

I analyzed the codebase and identified that the parseMentionsConfig() function in pkg/workflow/safe_outputs_config_messages.go was processing the allowed list without normalizing the '@' prefix. Here's what I fixed:

🔧 Code Changes

Modified: pkg/workflow/safe_outputs_config_messages.go

• Added normalization logic to detect and strip '@' prefix from usernames
• Added debug logging to track when normalization occurs
• Preserved usernames without '@' prefix unchanged

// Normalize username by removing '@' prefix if present
normalized := str
if len(str) > 0 && str[0] == '@' {
    normalized = str[1:]
    safeOutputMessagesLog.Printf("Normalized mention '%s' to '%s'", str, normalized)
}
allowedStrings = append(allowedStrings, normalized)

✅ Test Coverage

Modified: pkg/workflow/safe_outputs_mentions_test.go

• Added test cases for usernames with '@' prefix
• Added test cases for mixed formats (with and without '@')
• Verified end-to-end normalization through extractSafeOutputsConfig

All test cases pass, including edge cases like empty strings and single '@' characters!

How It Works

Before this fix:

mentions:
  allowed: ["`@pelikhan`"]  # ❌ Would fail at runtime

After this fix:

mentions:
  allowed: ["`@pelikhan`"]  # ✅ Normalized to "pelikhan" at compile time

Both formats now work seamlessly:

mentions:
  allowed: ["`@pelikhan`", "user2", "`@bot1`"]
  # Compiles to: ["pelikhan", "user2", "bot1"]

Benefits

🎨 User-Friendly: Users can specify usernames naturally with '@' prefix
⚡ Compile-Time: Normalization happens once during compilation, not at runtime
🔄 Backward Compatible: Existing configs without '@' continue to work
📝 Debug-Friendly: Logging helps track normalization during compilation
✨ Consistent: Matches the runtime sanitization expectations

Pull Request Created

I've created a pull request with:

• ✅ Clean, focused implementation
• ✅ Comprehensive test coverage
• ✅ Detailed documentation
• ✅ Verification of the fix

The PR is ready for review and includes all the changes described above with full test coverage!

---

Avec style et élégance 🎭✨

🎤 Magnifique! Performance by /cloclo for issue #15975

Warning

⚠️ Firewall blocked 1 domain

The following domain was blocked by the firewall during workflow execution:

• proxy.golang.org