Investigate and repair an agent's context — memory, system prompt, and skills — using observed behavior as evidence. Use when the user asks to audit, clean up, or fix your memory or context, or runs /doctor. Do not load for routine memory reads or edits, or requests that merely mention memory.
The skill exposes the agent to untrusted, user-generated content from public third-party sources, creating a risk of indirect prompt injection. This includes browsing arbitrary URLs, reading social media posts or forum comments, and analyzing content from unknown websites.
The workflow directs the agent to read historical conversation transcripts, messages, and chat history (`letta messages list`, `search`, `transcript`), which contain outsider-submitted free text from chat users.
letta-ai/letta-code
64fdf59
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.