Install and run the Continue CLI (`cn`) to execute AI agent checks on local code changes. Use when asked to "run checks", "lint with AI", "review my changes with cn", or set up Continue CI locally.
npxskills add continuedev/continue--skill cn-checkLoading…
Run AI-powered code checks locally against your working tree changes using the Continue CLI. Each check is an agent (defined in markdown) that reviews your diff, identifies issues, and optionally suggests fixes as a patch.
cn check in a projectnpm install -g @continuedev/cli
cn login
This opens a browser for authentication. After login, Hub-configured checks are available automatically.
cn check
This auto-detects checks from three sources (in priority order):
.continue/agents/*.md# Run a single local agent
cn check --agent .continue/agents/security-review.md
# Run a Hub-published agent
cn check --agent myorg/code-style
# Run multiple agents
cn check --agent .continue/agents/security.md --agent .continue/agents/docs.md
cn check --base develop
Default: auto-detects main or master.
# JSON output (for CI pipelines or scripting)
cn check --format json
# Unified patch output (pipe to git apply)
cn check --patch | git apply
# Stop on first failure
cn check --fail-fast
cn check --fix
Runs all checks, then applies any suggested patches directly to the working tree. Patches that conflict are reported but skipped.
Create a markdown file at .continue/agents/<name>.md:
# Security Review
You are a security reviewer. Examine the code changes for:
- SQL injection vulnerabilities
- XSS risks in user-facing output
- Hardcoded secrets or credentials
- Insecure use of eval() or similar
If you find issues, edit the files to fix them. If everything looks good, say so and exit.
The agent receives:
Any edits the agent makes are captured as a patch and reported as a "fail" with suggested changes.
git diff <base>...HEAD to find changed files--agent flags)git diff in the worktree as a patchChecks run in parallel by default. Use --fail-fast for sequential execution that stops on first failure.
A live-updating table shows check progress:
cn check - 3 checks against main - 5 changed files
Check Status Time
--------------------------------------------
Security Review * Running 12s
Code Style Pass 8s
Documentation Pending -
When complete, a full report prints with pass/fail status, agent output, and suggested patches.
--format json){
"checks": [
{
"agent": ".continue/agents/security.md",
"name": "security",
"status": "pass",
"patch": "",
"output": "No security issues found.",
"duration": 8.2
}
],
"summary": {
"total": 1,
"passed": 1,
"failed": 0,
"errored": 0
}
}
cn check [options]
Options:
--base <branch> Base branch for diff (default: auto-detect)
--format <format> Output format: text or json (default: text)
--fix Apply suggested fixes to working tree
--patch Output unified patch (pipe to git apply)
--fail-fast Stop after first failing check
--agent <agent> Agent to run (hub slug or local path, repeatable)
--config <path> Path to config file
--org <slug> Organization slug
--verbose Enable debug logging
| Problem | Solution |
|---|---|
| "No changes detected" | Make sure you have uncommitted changes or specify --base |
| "No checks found" | Create .continue/agents/*.md files or run cn login for Hub checks |
| Check times out (5 min) | Reduce diff size or split into focused agents |
| "Worker exited with code 1" | Run with --verbose to see worker stderr |
Patch conflicts with --fix | Apply patches manually: cn check --patch > changes.patch && git apply changes.patch |
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Set up and use 1Password CLI (op). Use when installing the CLI, enabling desktop app integration, signing in (single or multi-account), or reading/injecting/running secrets via op.
CLI to manage emails via IMAP/SMTP. Use `himalaya` to list, read, write, reply, forward, search, and organize emails from the terminal. Supports multiple accounts and message composition with MML (MIME Meta Language).
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Create or update AgentSkills. Use when designing, structuring, or packaging skills with scripts, references, and assets.
Set up and use 1Password CLI (op). Use when installing the CLI, enabling desktop app integration, signing in (single or multi-account), or reading/injecting/running secrets via op.
CLI to manage emails via IMAP/SMTP. Use `himalaya` to list, read, write, reply, forward, search, and organize emails from the terminal. Supports multiple accounts and message composition with MML (MIME Meta Language).