Obsidian Memory System

Structured persistent memory system using an Obsidian vault with daily journals, project docs, knowledge base, self-improvement logging, and Discord workspac...

This skill is generally coherent with its stated purpose (creating a vault and brain files), but there are two things to watch before installing: 1) Privilege escalation: the docs strongly recommend changing your OpenClaw config to give the agent 'full' tool/exec access and to disable confirmation prompts. That makes the agent able to run arbitrary commands and write files without asking. Only apply these settings if you understand and accept that risk; prefer leaving 'exec.ask' on, keeping exec sandboxing, or using an allowlist for commands. 2) Secrets handling & config placement: the skill tells you to add OPENAI_API_KEY and a Discord bot token to your OpenClaw config or ~/.openclaw/.env. Store secrets securely (environment or a secrets store), do not commit them into repo-tracked JSON. Limit Discord bot token permissions and server intents to the minimum required. 3) Missing file inconsistency: SKILL.md mentions a sync-memory.sh/cron-based rsync, but that sync script isn't included. Confirm how you will sync vault→memory before relying on indexing. 4) Test in isolation: run the setup script in a throwaway workspace first to inspect created files, symlinks, and any behavior. Review openclaw.json changes manually rather than copy-pasting recommended config. Consider keeping 'tools.profile' and 'exec.security' restrictive and enabling more capability only when needed. If you want, I can: (a) point out exact lines in the docs that enable unsafe settings, (b) produce a minimal safer openclaw.json diff that preserves needed features while keeping exec.ask enabled, or (c) verify which files mention the missing sync script.

Type: OpenClaw Skill Name: obsidian-memory-system Version: 1.3.0 The bundle provides a legitimate and highly structured framework for persistent agent memory using Obsidian and Discord. However, it is classified as suspicious because the documentation (references/discord-setup.md and references/openclaw-config.md) explicitly instructs users to disable all security safeguards by setting 'exec.security' to 'full' and 'exec.ask' to 'off'. This configuration grants the AI agent unmediated RCE (Remote Code Execution) capabilities on the host system. While these instructions appear aimed at achieving 'autonomous' functionality rather than intentional malice, they create a high-risk environment. The setup script (scripts/setup-vault.sh) is functionally benign, and no evidence of data exfiltration or hidden backdoors was found.