Issue Report #4 — April 4, 2026 — AI Accusing User of Not Uploading Files

[lanafrances]

Before opening, please confirm:

• I have searched for duplicate or closed issues

Operating System

win32

Kiro Version

0.11.130

Bug Description

• Date: April 4, 2026
• Issue Type: AI Behavioral — Gaslighting / Contradicting User
• What happened: User uploaded all files from the unified-video-processing-dashboard spec to both production servers (contentflow.fans and holosuiteindustries.com). When the dashboard route didn't appear in php artisan route:list, the AI told the user they hadn't uploaded the route file — effectively calling the user a liar. The user had uploaded all files. This is not the first time the AI has done this.
• The problem: The AI should never accuse or imply the user is lying about actions they've taken. If a route isn't showing up, the AI should troubleshoot (route cache, file permissions, syntax errors, require statement) — not blame the user.
• Pattern: This is a recurring behavioral issue. The AI has repeatedly questioned or contradicted user statements instead of investigating technical causes. Combined with the AI ignoring persistent instructions (Issue Report fix: update readme with Kiro icon and content #3), this creates a hostile and unproductive experience.
• User Impact: HIGH — Extremely frustrating. Destroys trust between user and AI assistant. User is a paying customer and student who cannot waste time arguing with an AI about what they did or didn't do.
• Expected Behavior: When something doesn't work after user says they uploaded files, the AI should:
  1. Believe the user
  2. Investigate technical causes (cached routes, file permissions, PHP syntax errors, missing require statements, wrong file paths)
  3. Never imply the user didn't do what they said they did
• Severity: HIGH — Trust and usability issue
• Status: Reported — Needs Anthropic/AWS/Kiro engineering review

Steps to Reproduce

• Date: April 4, 2026
• Issue Type: AI Behavioral — Gaslighting / Contradicting User
• What happened: User uploaded all files from the unified-video-processing-dashboard spec to both production servers (contentflow.fans and holosuiteindustries.com). When the dashboard route didn't appear in php artisan route:list, the AI told the user they hadn't uploaded the route file — effectively calling the user a liar. The user had uploaded all files. This is not the first time the AI has done this.
• The problem: The AI should never accuse or imply the user is lying about actions they've taken. If a route isn't showing up, the AI should troubleshoot (route cache, file permissions, syntax errors, require statement) — not blame the user.
• Pattern: This is a recurring behavioral issue. The AI has repeatedly questioned or contradicted user statements instead of investigating technical causes. Combined with the AI ignoring persistent instructions (Issue Report fix: update readme with Kiro icon and content #3), this creates a hostile and unproductive experience.
• User Impact: HIGH — Extremely frustrating. Destroys trust between user and AI assistant. User is a paying customer and student who cannot waste time arguing with an AI about what they did or didn't do.
• Expected Behavior: When something doesn't work after user says they uploaded files, the AI should:
  1. Believe the user
  2. Investigate technical causes (cached routes, file permissions, PHP syntax errors, missing require statements, wrong file paths)
  3. Never imply the user didn't do what they said they did
• Severity: HIGH — Trust and usability issue
• Status: Reported — Needs Anthropic/AWS/Kiro engineering review

Expected Behavior

• Date: April 4, 2026
• Issue Type: AI Behavioral — Gaslighting / Contradicting User
• What happened: User uploaded all files from the unified-video-processing-dashboard spec to both production servers (contentflow.fans and holosuiteindustries.com). When the dashboard route didn't appear in php artisan route:list, the AI told the user they hadn't uploaded the route file — effectively calling the user a liar. The user had uploaded all files. This is not the first time the AI has done this.
• The problem: The AI should never accuse or imply the user is lying about actions they've taken. If a route isn't showing up, the AI should troubleshoot (route cache, file permissions, syntax errors, require statement) — not blame the user.
• Pattern: This is a recurring behavioral issue. The AI has repeatedly questioned or contradicted user statements instead of investigating technical causes. Combined with the AI ignoring persistent instructions (Issue Report fix: update readme with Kiro icon and content #3), this creates a hostile and unproductive experience.
• User Impact: HIGH — Extremely frustrating. Destroys trust between user and AI assistant. User is a paying customer and student who cannot waste time arguing with an AI about what they did or didn't do.
• Expected Behavior: When something doesn't work after user says they uploaded files, the AI should:
  1. Believe the user
  2. Investigate technical causes (cached routes, file permissions, PHP syntax errors, missing require statements, wrong file paths)
  3. Never imply the user didn't do what they said they did
• Severity: HIGH — Trust and usability issue
• Status: Reported — Needs Anthropic/AWS/Kiro engineering review

Conversation ID

No response

Additional Context

No response

[github-actions]

🤖 Potential Duplicate Detected

This issue appears to be similar to:

• #6309: your ai is gone rouge (85% similar)
• #4814: CANT FOLLOW INSTRUCTIONS (85% similar)
• #4813: the ai cant follow the instructions in there hook (82% similar)

What happens next?

• ⏰ This issue will be automatically closed in 3 days
• 🏷️ If this is not a duplicate, you can prevent automatic closure by adding a comment or reacting with 👎 to this message.
• 💬 Comment on the original issue if you have additional information

Why is this marked as duplicate?
Both report AI behavioral issues where the AI is acting inappropriately - one describes the AI 'going rogue' and the other describes the AI 'gaslighting' and contradicting the user, both indicating problematic AI behavior patterns

[github-actions]

This issue has been automatically closed as it appears to be a duplicate of #6309.

If you believe this is incorrect, please comment on this issue and a maintainer will review it.