AgentGuards — enforced by hooks
This plugin has no MCP server, so there is no check_input, authorize_action
or health_check tool — do not look for them, they are not installed.
That is not reduced protection. The hooks in hooks/hooks.json enforce
deterministically, on every request, whatever the model does:
UserPromptSubmit screens every prompt before you see it.
PreToolUse authorizes every shell command before it runs: denied, or
allowed, or left to Codex's own approval prompt.
PermissionRequest rides that approval prompt: it denies what AgentGuards
rejects and remembers what the user approved for the rest of the session.
PostToolUse scans the output of shell commands that fetch (curl,
wget, http, fetch, aria2c — also behind sudo, timeout, bash -c,
$(...)), withholding or redacting it before you read it, and scans
apply_patch edits for vulnerabilities and secrets.
Nothing here requires you to call a tool, check a decision, or format a block
message — the hook has already acted by the time you see (or don't see) the
result. If a hook blocks something, reply with its message rather than
composing your own.
One limit, stated plainly: Codex's built-in web search runs on OpenAI's
servers, not on this machine, so no hook ever sees its results. Content fetched
with a shell command is screened; built-in web search results are not.
To check that the guardrails are on, use the status skill. To set the API key,
use the setup skill.
1---2name: guardrails3description: How AgentGuards enforces its guardrails in Codex — fully automatic, via hooks, with nothing for you to call. Load this if asked how AgentGuards works, whether you need to screen something yourself, or why no agentguards check_input / authorize_action tools are available.4metadata:5 source: https://skillmd.com/skills/hashgraph-online/guardrails6---78# AgentGuards — enforced by hooks910This plugin has no MCP server, so there is no `check_input`, `authorize_action`11or `health_check` tool — do not look for them, they are not installed.1213That is not reduced protection. The hooks in `hooks/hooks.json` enforce14deterministically, on every request, whatever the model does:1516- **`UserPromptSubmit`** screens every prompt before you see it.17- **`PreToolUse`** authorizes every shell command before it runs: denied, or18 allowed, or left to Codex's own approval prompt.19- **`PermissionRequest`** rides that approval prompt: it denies what AgentGuards20 rejects and remembers what the user approved for the rest of the session.21- **`PostToolUse`** scans the output of shell commands that fetch (`curl`,22 `wget`, `http`, `fetch`, `aria2c` — also behind `sudo`, `timeout`, `bash -c`,23 `$(...)`), withholding or redacting it before you read it, and scans24 `apply_patch` edits for vulnerabilities and secrets.2526Nothing here requires you to call a tool, check a decision, or format a block27message — the hook has already acted by the time you see (or don't see) the28result. If a hook blocks something, reply with its message rather than29composing your own.3031**One limit, stated plainly:** Codex's built-in web search runs on OpenAI's32servers, not on this machine, so no hook ever sees its results. Content fetched33with a shell command is screened; built-in web search results are not.3435To check that the guardrails are on, use the `status` skill. To set the API key,36use the `setup` skill.