Nexus Sentinel

ReviewAudited by ClawScan on May 10, 2026.

Overview

Nexus Sentinel is a coherent SRE automation skill, but it asks for broad server-recovery and cloud-backup authority without enough boundaries around restarts, cleanup, and sensitive-file handling.

Only install this if you intentionally want an agent to help administer the host. Before using it, define exactly which services and paths it may touch, require approval for restarts/cleanup/backups, use least-privilege Maton credentials, and verify the incomplete backup and safety enforcement code yourself.

Findings (4)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

A mistaken or overbroad action could remove useful files/logs or restart a production service, causing downtime or data loss.

Why it was flagged

The skill instructs the agent to perform host mutations such as cleanup/log rotation and automatic service restarts, but the artifacts do not define allowed services, allowed paths, rollback, or explicit confirmation for these high-impact actions.

Skill content
`/nexus optimize` : Rotation des logs et nettoyage des fichiers temporaires ... `/nexus recover <service>` : ... Service Restart ... `2 redémarrages automatiques par heure`
Recommendation

Use only with an explicit service allowlist and cleanup path list; require confirmation or dry-run for destructive actions and enforce restart limits in code.

What this means

A config file containing secrets could be backed up through the third-party Maton/GDrive flow without being detected by the shown check.

Why it was flagged

The backup gate shown before a cloud upload workflow relies on filename keyword matching, not content scanning or an allowlisted backup path, so sensitive configuration data could be treated as safe if the filename lacks those words.

Skill content
SENSITIVE_PATTERNS = [".env", "key", "password", "secret", "token"] ... if any(p in file_path.lower() for p in SENSITIVE_PATTERNS): ... # Logic to tar and push to GDrive via Maton ... return {"status": "ready_for_upload", "file": file_path}
Recommendation

Require explicit user-selected backup paths, content-based secret scanning, clear destination/retention settings, and per-backup approval before any cloud upload.

What this means

Reports may be sent to the configured WhatsApp destination using the provided Maton API key.

Why it was flagged

The skill uses a bearer token to send reports via the Maton WhatsApp gateway; this is aligned with the stated alerting purpose but grants access to an external account/channel.

Skill content
api_key = os.getenv("MATON_API_KEY") ... requests.post(f"{API_GATEWAY}/whatsapp/send", ... headers={"Authorization": f"Bearer {api_key}"})
Recommendation

Use a least-privilege API key, confirm the report channel, and avoid putting broad or reusable credentials in the environment.

What this means

A user may need to install local system tools or Python packages that were not declared in the registry requirements.

Why it was flagged

The dependencies are purpose-aligned for an SRE tool, but there is no install spec or pinned package versions, and the registry-level requirements shown for the skill are empty.

Skill content
Ce skill nécessite : `docker`, `pm2`, `tar`, `curl` ... `psutil` et `requests`. Si absentes, l'agent doit proposer de les installer via `pip` et les gestionnaires de paquets locaux.
Recommendation

Manually review and install dependencies from trusted sources, pin package versions where possible, and do not allow automatic package-manager changes without approval.