Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

Obsidian GitHub Sync

v1.0.0

Automated GitHub synchronization for Obsidian vault with conflict detection and notification. Use when the user wants to: - Sync their Obsidian vault to a Gi...

0· 94·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Benign
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
The name/description (Obsidian <-> GitHub sync) aligns with the included scripts: they commit local changes, pull --rebase, and push to a remote. The behavior (conflict flagging, cron/systemd examples) is consistent with the stated purpose. Minor mismatch: scripts assume branch name 'master' which may not match many repos using 'main'.
Instruction Scope
SKILL.md and the two scripts instruct the agent (and a user) to read and write the vault directory, create/modify the .git repository, write logs (/tmp/obsidian-sync.log) and a conflict flag (/tmp/obsidian-sync-conflict.flag), and run git operations (init, remote add, commit, pull --rebase, push). These actions stay within the scope of syncing but are far-reaching (they can initialize repos and modify git history). The instructions do not request or transmit data to any unexpected external endpoints beyond the configured GitHub remote.
Install Mechanism
No install spec and only small shell scripts are included. Nothing is downloaded from external URLs and no archives are extracted. Risk from installation is low because there is no automated network install step.
!
Credentials
The registry metadata lists no required environment variables, but SKILL.md and the scripts require OBSIDIAN_VAULT_DIR and GITHUB_REMOTE_URL (and optionally other env vars). That mismatch is an incoherence: the skill effectively needs filesystem and git/SSH access but the registry does not declare or surface those requirements. The requested env variables themselves are reasonable for the task (they don't include unexpected secret tokens), but GITHUB_REMOTE_URL implies use of SSH keys or credentials which the skill does not manage.
Persistence & Privilege
always:false and the skill is user-invocable; it does not demand forced always-on presence. The scripts suggest adding cron/systemd timers but that is optional and under user control. The skill does modify the vault directory (including initializing a .git repo) which is expected for its purpose but is a privileged filesystem action the user should approve.
What to consider before installing
This skill appears to do what it says (auto-commit, pull --rebase, push, and signal conflicts), but review and test before deploying. Concrete precautions: - Inspect the two scripts yourself. They run git init / git remote add and will modify the vault directory and .git metadata; run them first on a throwaway test copy of your vault. - The registry metadata does not declare required env vars; set and verify these before running: OBSIDIAN_VAULT_DIR and GITHUB_REMOTE_URL. Ensure the remote URL/branch is correct (script hardcodes 'origin master'). - The script relies on your Git/SSH credentials (it does not prompt for or store tokens). Ensure your SSH keys are configured and that you understand whether the remote uses 'main' vs 'master'. - Backup your vault before enabling automated cron/systemd runs. Automated commits + rebase can rewrite history or cause surprises if misconfigured. - Consider changing log/flag file locations and permissions if multiple users share the system; /tmp is world-readable on many systems and will contain repository path/remote in the flag file. - If you want the skill in an agent context, require that the agent only run it with an explicit configured OBSIDIAN_VAULT_DIR (avoid letting the skill infer paths) and avoid granting broad filesystem access. If you want, I can point out exact lines in the scripts to modify (branch name, logging path, preventing automatic git init/remote add) and produce a safer variant you can test.

Like a lobster shell, security has layers — review code before you run it.

latestvk97bp2az7d595waa1scp94m2e1836bv8

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments