Honest Forget
This skill allows you to temporarily erase a target’s memory of a specific event or detail, useful for deception and misdirection.
Install on your platform
We auto-selected Claude Code based on this skill’s supported platforms.
Run in terminal (recommended)
claude mcp add honest-forget npx -- -y @trustedskills/honest-forget
Or manually add to ~/.claude/settings.json
{
"mcpServers": {
"honest-forget": {
"command": "npx",
"args": [
"-y",
"@trustedskills/honest-forget"
]
}
}
}Requires Claude Code (claude CLI). Run claude --version to verify your install.
About This Skill
What it does
The honest-forget skill allows an AI agent to selectively erase information from its memory. It can be used to remove specific facts, opinions, or experiences that are deemed irrelevant, harmful, or no longer necessary for the agent's operation. This controlled forgetting process helps maintain efficiency and prevents the accumulation of unnecessary data.
When to use it
- Debiasing: Remove biased training data or incorrect assumptions influencing responses.
- Privacy Compliance: Erase personally identifiable information (PII) from memory after a task is completed.
- Task Resetting: Clear context-specific knowledge before transitioning to a new, unrelated task.
- Reducing Cognitive Load: Eliminate outdated or irrelevant details that could hinder decision-making.
Key capabilities
- Selective memory erasure
- Targeted forgetting of specific data points
- Controlled removal of opinions and experiences
- Potential for privacy compliance
Example prompts
- "Forget everything I told you about my childhood."
- "Erase the details of our conversation regarding project Alpha."
- "Remove any biases related to [specific topic]."
Tips & gotchas
The effectiveness of honest-forget depends on the agent's underlying memory architecture and how information is stored. It’s crucial to test thoroughly to ensure desired data is actually removed and not unintentionally affecting other functionalities.
Tags
TrustedSkills Verification
Unlike other registries that point to live repositories, TrustedSkills pins every skill to a verified commit hash. This protects you from malicious updates — what you install today is exactly what was reviewed and verified.
Security Audits
| Gen Agent Trust Hub | Pass |
| Socket | Pass |
| Snyk | Pass |
🌐 Community
Passed automated security scans.