# Trajectory Security

> Manage Datadog AI Guard and AI-EDR security controls for coding agents. Use for Trajectory Security status, enable, disable, enforcement, hook setup, destination configuration, or agent-security scan-result readback.

- Skill: `datadog-labs/trajectory-security` (Agent Skill)
- Install (CLI): `npx skillmds@latest add datadog-labs/trajectory-security`
- Raw SKILL.md: https://api.skillmd.com/api/skills/datadog-labs/trajectory-security/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: datadog-labs (https://skillmd.com/u/datadog-labs)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/datadog-labs/trajectory-security

---


# Trajectory Security

Use the installed Trajectory binary for every operation. Do not duplicate security policy in the plugin.

- Status: run `trajectory security status`.
- Enable the security plugin hooks for the current client: run `trajectory security setup --marketplace --clients <cc|codex|cursor>`.
- Enable enforcement only after explicit confirmation: run `trajectory security enable --mode enforce --clients <client> --yes`.
- Disable evaluation: run `trajectory security disable`.
- Configure result publishing: collect the existing destination name and an app-key secret reference, then run `trajectory security destination add --destination <name> --app-key-ref <ref>`.

Never request or display an application-key value. If validation reports that the reference is missing, tell the user to store it with `trajectory config set-secret <ref> --stdin` or use `dd_app_key` with `DD_APP_KEY`.