CLI reference

slopcop [flags]            # check the working tree (same as slopcop check)
slopcop [command] [flags]

Running slopcop with no subcommand runs the check. slopcop --help and slopcop <command> --help print the same text as this page. slopcop completion <shell> generates shell completion.

Global flags

Flag	Meaning
--service URL	the SlopCop check service URL (also SLOPCOP_SERVICE); default https://slopcop.com
-v, --version	print the version and exit
-h, --help	help for the command

slopcop check

Check the working tree through the hosted check API. See Run a check.

slopcop
slopcop check --format json
slopcop --root ../repo --base origin/main

Flag	Meaning
--root PATH	repository path to check (default .)
--base REV	branch or revision the change started from; the base is its merge base with HEAD (default: merge base with the upstream branch, then with origin/HEAD, then HEAD)
--base-exact REV	exact base commit or revision, used without taking a merge base
--format human|json|sarif	report format (default human)
--fail-on never|finding|note|warning|error	exit 1 when a new finding is at or above this level (default finding)
--policy PATH	repository-relative .rqlp policy file to evaluate (repeatable)
--no-policy-discovery	do not include .slopcop/policies rules automatically
--timeout DURATION	how long to wait for the result; a check still running then exits 3 (default 10m0s)
-w, --workspace NAME|ID	workspace to run in, and bill, by name or id, instead of the one this login is bound to (also SLOPCOP_WORKSPACE)
--add-repository	add the repository to the workspace --workspace names, which then pays for it, without asking at a terminal; pass it once the user has approved that
--verbose	print the resolved request to stderr, and name each inconclusive check in human

Exit codes: 0 clean, 1 new findings, 2 incomplete or failed, 3 still running, 4 refused and a person must act, 130/143 interrupted. See Exit codes.

slopcop login

Sign in to the SlopCop check service. With no flags, starts the browser device flow: prints a link that opens the approval page with the code filled in, then waits for approval. Also installs the agent skill into every well-known skills directory whose harness is installed here, as every check does. See Sign in.

Flag	Meaning
--token TOKEN	store an existing token instead of starting the browser flow

slopcop logout

Forget the stored credential.

slopcop whoami

Show the account behind this login, the workspace it is bound to (the one its checks and ignores run in, and bill), that workspace’s plan, and the service URL. Run it before a check when it matters which workspace pays.

Flag	Meaning
--format human|json	output format (default human); json has service, account, workspace, plan

slopcop workspace

List your account’s workspaces, or bind this login to another of them. See Workspaces.

Command	Meaning
workspace list [--format human|json]	the account’s workspaces; * marks this login’s
workspace use <name|id> [--format human|json]	bind this login to another workspace; moves every agent and script using the login, so it asks [y/N] and refuses when stdin is not a terminal

slopcop repo

List the repositories of this login’s workspace, or of the one --workspace names, or enable or disable one. See Repositories.

Command	Meaning
repo list [--format human|json]	the workspace’s repositories and whether each is active, disabled, or enabled over the limit
repo enable <owner/name|id>	enable a repository, within the workspace’s active limit (owner only; asks you to type the name)
repo disable <owner/name|id>	disable a repository: it is not checked or charged (owner only; asks you to type the name)

All take -w/--workspace. enable and disable take --format for their answer, and refuse when stdin is not a terminal.

slopcop ignore <finding-id>

Stop reporting one finding that is a false positive. The id is the one a check printed. Ignoring covers that one finding in its repository for your workspace, not its rule. Takes -w/--workspace. Exits 0 with a confirmation, 2 with the service’s message otherwise. See False positives.

slopcop init

Install the agent skill that ships inside this binary, and a Claude Code and Codex Stop hook that checks what a session wrote when the agent stops. See The agent skill and The Stop hook.

Flag	Meaning
--no-skill	do not install or update the agent skill
--no-hooks	do not install or update the automatic check hook
--remove-hooks	remove SlopCop’s automatic check hook and nothing else

slopcop hook <claude|codex>

Run a check from a Claude Code or Codex Stop hook. It reads the hook’s input on stdin and answers in the harness’s hook protocol. slopcop init installs it as slopcop hook claude || true and slopcop hook codex || true; it is not meant to be run by hand.

slopcop version

Print the version. A release binary prints its tag; a plain go install prints dev.

Environment variables

Variable	Meaning
SLOPCOP_TOKEN	a token for this run; wins over the stored credential
SLOPCOP_SERVICE	the service URL; --service wins over it
SLOPCOP_WORKSPACE	the workspace for check, ignore, and repo; -w wins over it
SLOPCOP_NO_UPDATE_NOTICE	any non-empty value silences the upgrade notice
CLAUDE_CODE_SESSION_ID, CLAUDE_CONFIG_DIR	set by Claude Code; name the session the CLI reads model attribution from
CODEX_THREAD_ID, CODEX_HOME	set by Codex; the same

See Configuration for precedence and the files the CLI writes.