Home / agent-security

Agent Security

Security skills for LLM agents: threat model a codebase, audit agent configuration, review MCP servers, trace prompt injection to tool calls, check consequential tools for human approval gates, look up incident precedents, run security evals, scan with Semgrep, and review agent session logs for injection, misuse, egress and loops. Includes guard hooks for Bash and an optional incident-database MCP server.

Plugin agent-security, version 0.6.0, 10 skills, MIT licence. Source: agent-security-skills. Synced .

Install

In Claude Code, add the marketplace and install the plugin:

/plugin marketplace add basitalisandhu/claude-skills
/plugin install agent-security@claude-skills

Or copy the skill files into ~/.claude/skills/ from a clone:

git clone https://github.com/basitalisandhu/claude-skills
cd claude-skills
python3 install.py --user --only agent-security

This plugin also ships plugin-level hooks, commands, agents and an MCP server. They load when you install the plugin from the marketplace; install.py copies skills only.

Skills

SkillWhat it doesScripts
agent-config-auditScan an AI agent configuration for risky permissions, leaked secrets, unpinned MCP servers and prompt injection in instruction files, with a bundled scanner that masks secrets.yes
agent-eval-harnessBuild a security evaluation for an agent with benign tasks, injections planted in tool results, and utility and attack success rate (ASR) scores, using a bundled standard-library runner and a policy hook.yes
agent-session-log-reviewReview an AI agent session log, transcript or chat export for signs that prompt injection took effect, tool misuse, destructive commands, denied actions retried another way, data leaving to unknown hosts and loops, with a bundled script that normalises the log and writes a timeline of flagged events.yes
agent-threat-modelWrite a threat model or threat assessment for an LLM-agent system by drafting principals, tools, channels, data stores and controls in YAML from the code, running the atm analyser for STRIDE and OWASP findings, and summarising with incident precedents.yes
incident-lookupLook up 80 coded AI-agent security incidents from 2023 to 2026 (prompt injection, tool poisoning, exfiltration and more), mapped to OWASP Agentic and LLM Top 10 and MITRE ATLAS techniques, and list real precedents with sources.yes
mcp-server-reviewReview an MCP server you wrote or plan to install for missing authentication, binding to a public interface, tool description poisoning, path traversal, SSRF, weak input validation and secrets in logs, then report a verdict per area with severity and file:line.no
prompt-injection-reviewTrace untrusted input such as web pages, emails, tickets and tool results to consequential tool calls in agent code (sending replies, HTTP requests, shell, payments), and judge each path against provenance and approval rules.yes
secure-agent-checklistRun a pre-ship security checklist for an LLM agent covering identity, least privilege, approvals, sandboxing, audit logging, kill switch, supply chain and evals, and report pass, fail or n.a. per item with evidence for a go or no-go sign-off.no
semgrep-agenticScan Python, JavaScript and TypeScript agent code with 36 Semgrep rules (model output flowing into shells, subprocess, SQL, paths or HTML, user input in system prompts, MCP servers without auth, leaked provider keys), with a bundled offline fallback, and triage the hits.no
tool-approval-gate-auditFind every tool a Python agent's model can call, decide which have consequential side effects (deleting files, sending email, cloud writes), and check whether a human approves each one through a deterministic gate that cannot fail open, with gaps reported per tool at file and line.yes