Skip to main content
Use aislop commands in your terminal for the same command list. Use aislop <command> --help for command-specific help.

Command map

CommandWhat it does
aislop [directory]Opens the interactive menu in a TTY, or scans the current directory in non-TTY shells
aislop scan [directory]Scores code quality and prints findings
aislop agent [directory]Runs a local worktree repair session with Codex, Claude Code, or OpenCode
aislop agent plan [directory]Previews provider, worktree, findings, and publish actions without editing
aislop agent providersShows installed local provider status and setup hints
aislop agent connect [provider]Runs a provider’s local CLI login flow
aislop agent use [provider]Sets or shows the repo-local default repair provider
aislop agent switch [provider]Alias for aislop agent use
aislop agent monitor [directory]Watches local git changes and streams scan or repair cycles
aislop agent monitor list [directory]Lists local background agent monitors
aislop agent monitor show [monitor]Shows one background monitor record
aislop agent monitor stop [monitor]Stops a running background monitor
aislop agent sessions [directory]Lists recent local agent sessions
aislop agent show [session]Shows a local agent session summary and timeline
aislop agent apply [session]Applies a reviewed isolated worktree session back to the repo
aislop agent watch [session]Watches a local agent session as it streams
aislop agent stop [session]Stops a running background agent session
aislop fix [directory]Applies deterministic auto-fixes or prints/opens an agent handoff prompt
aislop ci [directory]Runs the CI quality gate with thresholded exit codes
aislop init [directory]Creates config, rules, and optional GitHub Actions workflow
aislop doctor [directory]Checks installed engines and project coverage
aislop rules [directory]Explains rule IDs, severity, fixability, score impact, and meaning
aislop hookManages per-edit coding-agent hooks
aislop hook install [agents...]Installs coding-agent hooks
aislop hook uninstall [agents...]Removes coding-agent hooks
aislop hooksAlias for aislop hook
aislop hook statusShows installed hook status
aislop hook baselineCaptures the current score as the hook baseline
aislop install [agents...]Alias for aislop hook install
aislop install hooks [agents...]Natural alias for aislop hook install
aislop uninstall [agents...]Alias for aislop hook uninstall
aislop uninstall hooks [agents...]Natural alias for aislop hook uninstall
aislop badge [directory]Prints score badge URL and README markdown
aislop trend [directory]Shows recent local scores from .aislop/history.jsonl
aislop trends [directory]Alias for aislop trend
aislop updateShows current and latest npm versions
aislop upgradeAlias for aislop update
aislop versionPrints the installed version
aislop commandsShows the complete command reference

Global help and version flags

FlagDescription
-h, --helpShows help
-v, --versionShows the installed version
-VShows the installed version

Scan-style flags

These flags are accepted by aislop [directory], aislop scan, and, where noted, aislop ci.
FlagCommandsDescription
--changesscan, ci, root commandScans or gates changed files
--stagedscan, ci, root commandScans or gates staged files
--base <ref>scan, ci, root commandDiff base for --changes, such as origin/main or FETCH_HEAD
-d, --verbosescan, root commandShows file details per rule
--jsonscan, root commandEmits machine-readable JSON
--sarifscan, ci, root commandEmits SARIF 2.1.0 for code scanning
--format <format>scan, ci, root commandChooses json or sarif
--include <patterns>scan, root commandRestricts scanning to comma-separated or repeated paths
--exclude <patterns>scan, root commandExcludes comma-separated or repeated paths
--humanciRenders the human-friendly scan UI instead of JSON

Fix flags

FlagDescription
-d, --verboseShows detailed fix progress
-f, --forceRuns aggressive fixes such as audits and framework dependency alignment
--safeOnly applies reversible fixes: imports, comment removal, and formatting
-p, --promptPrints a prompt for a coding agent to fix remaining issues
--claudeOpens Claude Code with remaining findings
--codexOpens Codex CLI with remaining findings
--cursorOpens Cursor and copies the prompt to clipboard
--windsurfOpens Windsurf and copies the prompt to clipboard
--vscodeOpens VS Code and copies the prompt to clipboard
--ampOpens Amp with remaining findings
--antigravityOpens Antigravity with remaining findings
--deep-agentsOpens Deep Agents with remaining findings
--geminiOpens Gemini CLI with remaining findings
--kimiOpens Kimi Code CLI with remaining findings
--opencodeOpens OpenCode with remaining findings
--warpOpens Warp with remaining findings
--aiderOpens Aider with remaining findings
--gooseOpens Goose with remaining findings
--piOpens pi with remaining findings
--crushOpens Crush with remaining findings

Agent flags

FlagDescription
--provider <provider>Provider to use: auto, codex, claude, or opencode
--target-score <score>Score to converge toward, default 90
--max-turns <n>Maximum provider turns for one repair attempt, default 4
--limit <n>Maximum findings to hand to the provider, default 8
--in-placeEdits the current worktree instead of creating an isolated worktree
--applyApplies the accepted diff back to the original worktree
-y, --yesSkips confirmation prompts for --apply
--dry-runPrints the selected provider and plan without running it
--backgroundStarts the agent in the background and returns immediately
--no-fixSkips deterministic safe fixes before provider handoff
--commitCommits the verified diff on an agent branch
--prPushes the agent branch and opens a draft pull request
--branch <name>Branch name for --commit or --pr
--base <branch>Base branch for --pr
--commit-message <message>Commit message for --commit or --pr
--title <title>Pull request title for --pr
--readyOpens a ready-for-review PR instead of a draft
--no-keep-worktreeRemoves the generated worktree when it is safe
--cleanupRemoves the generated worktree even when a diff remains

Agent monitor flags

FlagDescription
--provider <provider>Provider for monitor repair sessions: auto, codex, claude, or opencode
--target-score <score>Score monitor repair tries to converge toward
--max-turns <n>Maximum provider turns per repair attempt
--limit <n>Maximum findings to hand to the provider
--in-placeRequired with --repair; edits the current checkout
--dry-runPrints the monitor plan without starting it
--no-fixSkips deterministic safe fixes before provider handoff
--repairRuns bounded repair sessions when scans miss the target
--backgroundStarts a detached local monitor
--interval <ms>Poll interval for git status
--debounce <ms>Wait time after changes settle
--onceRuns one monitor cycle and exits
--root <directory>Root directory for monitor show/stop commands
--forceUses SIGKILL when stopping a monitor

Agent utility flags

CommandFlagDescription
agent connect--dry-runPrints the provider login command without running it
agent use, agent switch--root <directory>Git repository that stores the repo-local provider preference
agent use, agent switch--dry-runPrints the provider preference change without writing
agent monitor list--limit <n>Maximum monitor records to show
agent monitor show, agent monitor stop--root <directory>Git repository to read monitor records from
agent monitor stop--forceSends SIGKILL instead of SIGTERM
agent sessions--limit <n>Maximum session records to show
agent show--root <directory>Git repository to read sessions from
agent apply--root <directory>Git repository to read sessions from
agent apply--dry-runPreviews the session diff without applying it
agent apply-y, --yesSkips confirmation prompts
agent watch--root <directory>Git repository to read sessions from
agent watch--interval <ms>Poll interval while following the transcript
agent watch--oncePrints current session events and exits
agent stop--root <directory>Git repository to read sessions from
agent stop--forceSends SIGKILL instead of SIGTERM

Hook flags

FlagDescription
--agent <names>Comma-separated agent list
-g, --globalInstalls or uninstalls from user-scope config
--projectInstalls to project-scope config
--dry-runPrints the planned diff without writing
--yesSkips the confirmation prompt where supported
--quality-gateAdds a Claude Stop hook that blocks score regressions below baseline
--claudeShortcut for --agent claude
--cursorShortcut for --agent cursor
--geminiShortcut for --agent gemini
--piShortcut for --agent pi
--codexShortcut for --agent codex
--windsurfShortcut for --agent windsurf
--clineShortcut for --agent cline
--kilocodeShortcut for --agent kilocode
--antigravityShortcut for --agent antigravity
--copilotShortcut for --agent copilot

Reporting flags

CommandFlagDescription
badge--owner <owner>GitHub owner for badge generation
badge--repo <repo>GitHub repo for badge generation
badge--jsonEmits badge data as JSON
trend, trends--limit <n>Number of recent local score records to show

Setup and rule flags

CommandFlagDescription
init--strictWrites an enterprise-grade default config with all engines, typecheck on, ci.failBelow: 85, and workflow included
rules--searchOpens an interactive searchable rule explorer

Scope files

FilePurpose
.aislop/config.ymlProject policy: engines, scoring, thresholds, telemetry, include/exclude
.aislop/rules.ymlOptional Architecture engine rules
.aislopignoreProject-wide scan ignores
.gitignoreRespected for untracked files
.aislop/history.jsonlLocal score history written by interactive full scans
.aislop/agent/sessions/Local agent transcripts
For day-to-day use, start with aislop scan, repair with aislop fix --safe and aislop agent, then enforce the result with aislop ci.