Anomaly detection for AI agent behavior. Monitors tool call patterns, file access sequences, and command history to flag unusual activity — first-seen patterns, frequency spikes, scope creep, and suspicious sequences. Ported from Flare's cloud audit log anomaly detection engine. Trigger phrases: "watch my agent", "behavior audit", "what has the agent been doing", "session activity report", "anomaly check", "agent activity".
EnglishPer-session cost and token budget tracking for AI coding agents. Monitors cumulative spend, warns at thresholds, and enforces limits with three modes: reject (stop work), alert (warn and continue), downgrade (suggest cheaper approach). Ported from the LOCO-Agent BudgetManager. Trigger phrases: "set a budget", "track my spend", "cost limit", "how much have I spent", "token budget", "spending cap".
EnglishSafety check before executing destructive or irreversible commands. Catches dangerous shell commands, risky git operations, secret exposure, and high-blast-radius actions before they run. Activates automatically when Claude is about to execute shell commands that match known risk patterns. Trigger phrases: "check before running", "is this command safe", "safety check", "pre-execution review".
EnglishSecurity audit for AI agent skills before installation. Scans SKILL.md files, hooks, scripts, and MCP configs for prompt injection, data exfiltration, credential theft, and malicious automation patterns. Use when installing a new skill, reviewing a skills repo, or auditing existing installed skills. Trigger phrases: "verify this skill", "is this skill safe", "audit skill", "check before installing", "scan this plugin".
EnglishAllowlist/denylist for AI agent tool calls with approval gates for destructive operations. Define which tools the agent can use freely, which require confirmation, and which are blocked entirely. Combines LOCO-Agent's resource authorization with Flare's anomaly-aware security heuristics. Trigger phrases: "restrict tools", "tool permissions", "block bash", "approval gate", "limit what you can do", "tool whitelist", "tool blacklist".
English
AGENT SKILLS ATLASIncludes Behavior Watch, Cost Guard, Pre Exec Check, and other workflows.
Claude Code
Codex
Cursor
GitHub Copilot
Gemini CLISupported by the Atlas install prompt. Skill-specific requirements vary.
Pack stats
What’s included
Behavior Watch
Anomaly detection for AI agent behavior. Monitors tool call patterns, file access sequences, and command history to flag unusual activity — first-seen patterns, frequency spikes, scope creep, and suspicious sequences. Ported from Flare's cloud audit log anomaly detection engine. Trigger phrases: "watch my agent", "behavior audit", "what has the agent been doing", "session activity report", "anomaly check", "agent activity".
Cost Guard
Per-session cost and token budget tracking for AI coding agents. Monitors cumulative spend, warns at thresholds, and enforces limits with three modes: reject (stop work), alert (warn and continue), downgrade (suggest cheaper approach). Ported from the LOCO-Agent BudgetManager. Trigger phrases: "set a budget", "track my spend", "cost limit", "how much have I spent", "token budget", "spending cap".
Pre Exec Check
Safety check before executing destructive or irreversible commands. Catches dangerous shell commands, risky git operations, secret exposure, and high-blast-radius actions before they run. Activates automatically when Claude is about to execute shell commands that match known risk patterns. Trigger phrases: "check before running", "is this command safe", "safety check", "pre-execution review".
Skill Verify
Security audit for AI agent skills before installation. Scans SKILL.md files, hooks, scripts, and MCP configs for prompt injection, data exfiltration, credential theft, and malicious automation patterns. Use when installing a new skill, reviewing a skills repo, or auditing existing installed skills. Trigger phrases: "verify this skill", "is this skill safe", "audit skill", "check before installing", "scan this plugin".
Tool Guard
Allowlist/denylist for AI agent tool calls with approval gates for destructive operations. Define which tools the agent can use freely, which require confirmation, and which are blocked entirely. Combines LOCO-Agent's resource authorization with Flare's anomaly-aware security heuristics. Trigger phrases: "restrict tools", "tool permissions", "block bash", "approval gate", "limit what you can do", "tool whitelist", "tool blacklist".
Quick start
- Choose your skills. Select individual skills above, or choose the entire pack.
- Copy the install prompt. Use Copy install prompt above the list.
- Send it to your agent. Paste into your coding agent’s chat. Your agent reviews the source and guides the setup.
Try it with your agent
After installation, start with a request like:
Use the Behavior Watch skill to help with my task.
Ask me for the context you need before getting started.Requirements & license
- A coding agent that can access your project files and install skills.
- Your agent checks each skill’s instructions, dependencies, and permissions during setup.
- Reported license for the skills shown: MIT. The author’s terms apply.