More screenshots
Works with the AI Agents you already use
Database Security: Authorized Assessment Playbook
Assess database security with authorization across Postgres, MySQL, MSSQL, MongoDB, and Redis — every critical finding validated as an exploit chain, not a config observation.
$5
Database Security: Authorized Assessment Playbook
Example session with this skill installed
Assess our staging PostgreSQL database. The app account is app_rw, there is an old account legacy_etl nobody owns, and audit logging is off. What do you find and how do you report it?
- Read your context and instructions
- Compiled the database security
Assessment — staging PostgreSQL 15 (read-only checks)
Exposure: bound to localhost, TLS enforced, no public access — clean. Authentication: app_rw uses a 32-character generated password, rotated quarterly — clean. But legacy_etl exists with a password set in 2021 and no owner; its password sits in a shared runbook six former contractors could have seen.
Authorization: app_rw holds SELECT on the whole payments schema, broader than its queries need. legacy_etl holds superuser. Dangerous features: COPY ... PROGRAM is available to superusers — and legacy_etl can use it.
Validated chain (not just a configuration observation): from the runbook password, an attacker authenticates as legacy_etl, reaches superuser, and gets command execution via COPY ... PROGRAM. That is the critical finding.
Report: CRITICAL — orphaned superuser with shared stale password and command-execution path. Remediation: drop legacy_etl, rotate the runbook, restrict app_rw to the tables its queries touch. Separate hardening note: enable statement-level audit logging (currently off) — important, but not this week's fix. No destructive statements were run; nothing was written to the database. Rubric score: 7/8.
Connects securely to your tools. The creator never sees your data.
About this skill
Databases hold the crown jewels and are routinely the least hardened part of the stack. This skill runs an authorized assessment in a fixed order: network exposure (bind addresses, TLS, cloud public-access settings), authentication (defaults, weak passwords tested on non-production first, orphaned and shared accounts), authorization (role-to-grant mapping, who can read PII and payment data), dangerous features per engine (from xp_cmdshell to COPY ... PROGRAM to Redis CONFIG rewrite), injection paths through the application, and — the step that separates this from a checklist — safe validation of the exploitation chain for every high-severity finding.
Production is read-only by default, and the report separates what to fix this week (exploitable) from what to fix this quarter (hardening). A worked example walks a staging PostgreSQL instance to a validated critical: an orphaned superuser with a shared stale password and a command-execution path. A rubric gates the report: below 6 out of 8, the validation is not finished.
- It assesses; it does not fix. Remediation is specified per finding, but applying it is your change process.
- A finding without a validated path is reported as a hardening note, not a critical — the skill is strict about that distinction on purpose.
- It assumes you have authorized access and the right privilege level to inspect grants and configuration; read-only accounts may not see everything.
- Cloud console checks depend on console access; without it, the most common exposure class goes unchecked and the report says so.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 3 days ago
Permissions
No special permissions declared or detected