llm-wiki

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill includes an explicit command that passes a user password on the command line ("ob login --email --password ''"), which requires the agent to handle and potentially emit the secret value verbatim.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). The skill’s required “Ingest” workflow reads outsider-provided sources at runtime (URLs/files/pasted text), extracts their contents into raw/ (e.g., via npx defuddle <url> / web_extract) and then uses that readable text to write wiki pages—so outsider-authored free text can enter the LLM context indirectly through source ingestion.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The skill instructs creating and enabling a systemd service and explicitly runs a sudo command (sudo loginctl enable-linger $USER) and global installs (npm -g), which require elevated privileges and modify system-level state.