syncing-memory-filesystem

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The prompt includes explicit commands that embed API keys into headers and git credential helpers (e.g., Authorization: Bearer $LETTA_API_KEY and password=$LETTA_API_KEY in a single-line helper), which encourages producing or substituting secret values verbatim into generated commands — an exfiltration risk.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill explicitly clones/pulls the agent's memory git repo at "$LETTA_BASE_URL/v1/git//state.git" at runtime and injects the fetched Markdown files from memory/system/ into the agent's system prompt, so remote content directly controls agent prompts.