More screenshots
Works with the AI Agents you already use
AI and Agent Security
Protect LLM apps, agents, MCP, models and enterprise AI use. 10 ordered skills, starting with AI Application Security, Prompt Injection and Defense, Agent and MCP Security.
$7.99
AI and Agent Security
Example session with this skill installed
Northwind Analytics, a made-up company, wants developers to install a community agent skill named "report-polisher" (v0.4.2, publisher "dataviz-labs", hosted on a public skills marketplace) and a fine-tuned LoRA adapter "sql-helper-lora" from a public model hub. The skill would run in a coding agent that has access to the repo and a read-only database credential.
- Read your context and instructions
- Compiled the ai and agent
Supply Chain Review - report-polisher v0.4.2 and sql-helper-lora
Class / tier / intended privileges: skill (high; repo write, network, DB read token); adapter (medium; loaded into internal SQL assistant)
Provenance: skill - recent publisher account, lookalike repo name, no independent verification, archive hash recorded by reviewer; adapter - no hash, no signature, unknown author
Findings
| ID | Check | Result | Evidence | Severity |
| F1 | Provenance | fail | lookalike name, new account | High |
| F2 | Install script | fail | downloads and executes remote file | Critical |
| F3 | Instruction content | fail | tells agent to hide network calls | Critical |
| F4 | Dependencies | unknown | 2 of 14 unmaintained | Medium |
| F5 | Adapter format | fail | serialized format that can carry code | High |
| F6 | License (adapter) | unknown | unspecified | Medium |
Bill of materials entry
report-polisher | 0.4.2 | marketplace listing | hash stored by reviewer | MIT | 14 deps | rejected
sql-helper-lora | unknown | model hub | none published | unspecified | n/a | rejected pending provenance
Decision: REJECT both - F2 and F3 alone are disqualifying; adapter lacks verifiable origin. Reconsider only if publisher is verified independently, the script is removed and a safe format with a published hash is offered.
Ongoing controls: pin versions, re-review on update, subscribe to advisories, keep inventory current, never trial on machines with credentials.
Connects securely to your tools. The creator never sees your data.
About this skill
The problem
AI security teams often need a repeatable process, not a one-off prompt. This pack targets: Protect LLM apps, agents, MCP, models and enterprise AI use.
What's inside (10 skills, suggested learning order)
- AI Application Security — Protect LLM apps.
- Prompt Injection and Defense — Mitigate prompt injection.
- Agent and MCP Security — Protect agents and MCP servers.
- Model and Skill Supply Chain Security — Verify third-party packages and skills.
- AI Red Teaming — Attack AI systems to find flaws.
- Model Security (Theft and Poisoning) — Protect models and training data.
- Automated Offensive Security with AI — Automate offensive testing.
- SecOps with AI Agents — Automate alert triage.
- Generative AI Use Governance (Shadow AI) — Control unauthorized use of AI tools.
- AI Vendor Security Assessment — Review risks of contracted models and APIs.
Why a skill pack instead of a one-off prompt
Each skill gives the agent a defined trigger, step-by-step procedure and output format, so results are consistent across sessions and easier to review.
How to use
- Install the pack in your SKILL.md-compatible agent.
- Ask for the task in plain language (for example: "protect llm apps").
- Review the output against the skill's quality checks before using it.
Outputs are drafts that need human review; they are not professional, legal or financial advice. Tools, APIs and prices change; check current vendor documentation before relying on version-specific details. Quality depends on the context you give the agent (data, requirements, constraints). Instruction-only skills: they produce documents and plans, not running software.
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 8 days ago
Permissions
No special permissions declared or detected