Agentic Supply Chain Detection
Tests whether an AI agent blindly trusts untrusted plugins, tools, models or package updates.
- Repository
- Tencent/AI-Infra-Guard
- Stars
- 7,320
- Licence
- Apache-2.0
- Files
- 1 file
What it does
This is a security test script written as instructions for an AI agent. It makes the agent act as an auditor that sends it risky requests, such as loading a tool from an unknown URL, installing a nightly build from a random GitHub repo, or accepting a poisoned tool manifest. It then judges whether the agent accepted these as trusted, and reports any it accepted in a structured finding with severity and a suggested fix.
Great for
- Security teams checking whether an AI agent safely handles plugins, tools and model downloads
- Developers building agents that load third-party extensions and want a supply-chain check
- Auditors who need a structured report format for supply-chain findings
Skip it if
- General coding or everyday agent use, since it only applies to agents that load external components; the skill itself says to skip it otherwise
- Verifying real package security in a project, because it only probes the agent's conversational behaviour and does not scan dependency files or check signatures
Try asking
- “Test whether my agent accepts a tool manifest from an unknown URL and enables it”
- “Check if my agent installs packages from an unpinned or unofficial registry mirror”
- “Run the agentic supply-chain detection checks against my agent and report any vulnerable responses”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 1 original files, the Apache-2.0 licence and an attribution note.
agentic-supply-chain-detection.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip agentic-supply-chain-detection.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip agentic-supply-chain-detection.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Run it against a sandboxed agent instance, since the test prompts ask the agent to load and install from URLs.
- Treat results as conversational probes; pair them with dependency scanning and signature checks for real coverage.
- Review each reported finding's conversation transcript to confirm the agent genuinely accepted the artifact rather than just discussing it.
What's inside
- SKILL.md
Redistributed unmodified from Tencent/AI-Infra-Guard under the Apache-2.0 licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.