CLI Command Reference

Global configuration

Config lives at ~/.entelligence/config.json (created automatically, permissions locked to 0600 — owner read/write only). It stores your API key, org UUID, and GitHub token as they're set via auth login and related commands.

Resolution order (highest priority first) for any credential: CLI flag → matching environment variable → value saved in the config file.

Environment variable	Overrides	Notes
`ENTELLIGENCE_API_KEY`	`api_key` in config	Alternative to `entelligence auth login` — set this for CI/non-interactive use
`ENTELLIGENCE_TOKEN`	`--token` on `review`	Distinct from `ENTELLIGENCE_API_KEY` — used specifically by `review`'s `--token` flag
`ENTELLIGENCE_ORG_UUID`	`org_uuid` in config / `--org-uuid`	Also accepted by `entelligence incident --org-uuid`
`ENTELLIGENCE_ENDPOINT`	`--endpoint` on `review`	API endpoint override, distinct from `ENTELLIGENCE_BASE_URL` below
`ENTELLIGENCE_BASE_URL`	default API host	Production default: `https://entelligence.ddbrief.com`. Set to `http://localhost:8000` for local backend development
`ENTELLIGENCE_APP_URL`	default web app host	Used for the browser-based login flow. Default: `https://app.entelligence.ai`
`ENTELLIGENCE_GITHUB_TOKEN` / `GITHUB_TOKEN` / `GH_TOKEN`	`github_token` in config	Checked in that order
`ENTELLIGENCE_NO_AUTO_UPDATE`	—	Set to disable the daily auto-update check

Auto-update also skips automatically in CI — detected via CI, GITHUB_ACTIONS, GITLAB_CI, BUILDKITE, JENKINS_URL, or TF_BUILD being set — and for editable/dev installs, where it prints a one-line notice instead of upgrading in place.

Exit codes: every command exits 1 on any failure (auth error, network error, git error, invalid input) and 0 on success — there's currently no distinct exit code to distinguish "ran fine, found issues" from "ran fine, found nothing." Script around output content (or --json where available) rather than the exit code alone if you need to branch on findings.

Auth

Command	Description
`entelligence auth login`	Authenticate — paste your API key, or opens a browser-based login against `ENTELLIGENCE_APP_URL`
`entelligence auth status`	Show current authentication status
`entelligence auth logout`	Clear stored credentials

Review

Runs AI code review against your local git diff. See Reviewing from the CLI for a walkthrough.

Flag	Default	Description
`--base-branch`	`main`	Base branch to diff against
`--endpoint` (env `ENTELLIGENCE_ENDPOINT`)	from config	Override the API endpoint for this run
`--token` (env `ENTELLIGENCE_TOKEN`)	from config	Override the auth token for this run
`--org-uuid` (env `ENTELLIGENCE_ORG_UUID`)	resolved server-side	Override the org UUID for this run
`--priority`	`medium`	`low` / `medium` / `high` — how aggressively findings are surfaced
`--mode`	`verbose`	`concise` / `verbose`
`--plain`	off	Plain, detailed text output — no rich terminal formatting
`--prompt-only`	off	Minimal, token-efficient output, meant for piping into another agent
`--repo-path`	`.`	Path to the repo to review
`--committed-only`	off	Force a committed-vs-base-branch diff instead of trying uncommitted changes first
`--textual`	off	Experimental. Opens an interactive Textual file browser before reviewing
`--mock`	off	Exercise the UI with fake data — no API call, useful for scripting/demos
`--debug`	off	Print the raw request/response and write a log to `~/.entelligence/cli-debug.log`

entelligence review --priority high --mode verbose
entelligence review --committed-only --base-branch develop
entelligence review --prompt-only   # feed straight into another agent

`diff` and `status`

Command	Flags	Description
`entelligence diff`	`--file/-f <path>` (required), `--base-branch`, `--repo-path`	Show the diff for a single file
`entelligence status`	`--base-branch`, `--repo-path`	Repo and auth status summary

Security

Vulnerability scanning for a connected repo. Every subcommand supports --json for scripting.

Command	Flags	Description
`entelligence security scan <owner/repo>`	`--branch`, `--force`, `--wait` (polls every 15s, up to 10 min), `--json`	Kick off a scan
`entelligence security status <scan_id>`	`--json`	Check progress (`scan_status`, `progress_percentage`, `current_step`)
`entelligence security list [<owner/repo>]`	`--limit` (20), `--status-filter`, `--json`	List recent scans
`entelligence security summary <scan_id>`	`--json`	Score + severity breakdown
`entelligence security issues <scan_id>`	`--severity` (repeatable), `--type` (repeatable), `--file`, `--rule`, `--limit` (50), `--offset` (0), `--json`	List individual findings, filterable
`entelligence security suggest <scan_id> <rule_key> <owner/repo>`	`--file`, `--json`	Get a suggested fix for one finding
`entelligence security dashboard`	`--scan-id`, `--json`	Current metrics, severity breakdown, vulnerability trend

entelligence security scan acme/payments-service --wait
entelligence security issues <scan_id> --severity critical --severity high --json

💡

security doesn't accept a global --org-uuid override — org is always resolved from your authenticated session.

Incident

AI-driven triage and remediation for recurring production error patterns ("Error Knowledge Units" / EKUs). See the /incident skill for an interactive workflow.

Command	Flags	Description
`entelligence incident list`	`--status` (`active`/`dormant`/`all`, default `all`), `--repo`, `--verbose`, `--limit` (50)	Ranked list of error patterns by severity, then frequency
`entelligence incident show <eku_id>`	—	Root cause and recommended fix for one pattern
`entelligence incident logs <eku_id>`	`--limit` (20), `--full`, `--json`	Raw, deduplicated production log samples
`entelligence incident fix <eku_id> --repo <org/name>`	`--watch` (polls every 15s, up to 5 min), `--yes/-y`	Trigger an automated fix — opens a PR
`entelligence incident monitor`	`--phase`, `--repo`, `--verbose`	List merged fixes currently being verified post-merge
`entelligence incident track <job_id>`	`--watch`	Status of a fix job — classification, PR link, test results
`entelligence incident refresh`	`--verbose`	Trigger a fresh monitoring sweep

The whole group also accepts --org-uuid to target a specific org.

⚠️

fix is a write action — it opens a PR against the repo you specify. Unless you pass --yes, it prints the pattern, severity, and repo, then prompts "Trigger Fix Now?" (defaults to No — you must explicitly confirm).

entelligence incident list --status active --repo acme/payments-service
entelligence incident fix EKU-1842 --repo acme/payments-service --watch

Insights

Coding-agent usage analytics — see Agent Insights setup for the full walkthrough. Most subcommands are a thin passthrough to the open-source cinsights engine; sync and schedule are native to entelligence-cli.

Command	Key flags	Description
`entelligence insights setup`	`--provider`, `--model`, `--base-url`, `--extra-headers`, `--digest`, `--validate`	Configure the LLM used for local analysis. Omit provider/model to get an interactive prompt
`entelligence insights index`	`--source` (`local`/`entireio`/`phoenix`), `--hours` (default 4380 ≈ 6 months), `--force`, `-v`	Index local session files — free, no LLM calls
`entelligence insights analyze`	`--limit` (50), `--concurrency` (5), `--min-score` (0.0), `--yes/-y`, `--tasks-only`	Analyze indexed sessions. Prints a cost estimate and asks to confirm unless `--yes`
`entelligence insights digest project <name>`	`--days`	Generate a project digest
`entelligence insights digest user <id>`	`--days`	Generate a personal digest
`entelligence insights serve`	—	Local web dashboard at `http://localhost:8100`
`entelligence insights sync`	`--hours` (168), `--all`, `--no-index`, `--batch-size` (100), `--dry-run`	Push locally-computed metrics to your org's Agent Insights dashboard
`entelligence insights schedule`	`--at HH:MM` (default `09:00`), `--remove`	Install/remove a daily job that runs `insights sync` automatically

⚠️

Naming gotcha: cinsights (the underlying open-source tool) has its own schedule install/status/uninstall subcommand for scheduling local index+analyze. When invoked through entelligence insights schedule, the literal word schedule is intercepted by Entelligence's own native command first — so entelligence insights schedule install does not behave like cinsights schedule install. Only --at and --remove are supported through the entelligence wrapper.

What gets synced, and what doesn't: insights sync only pushes pre-computed metrics — token counts, tool-call names and counts, efficiency scores. It explicitly excludes raw context blobs and never sends tool input/output (which could contain code, commands, or secrets).

entelligence insights setup
entelligence insights index --source local --hours 8760
entelligence insights sync --dry-run   # see what would be pushed, without sending it
entelligence insights schedule --at 09:00

Wrapped

Command	Flags	Description
`entelligence wrapped`	`--days`, `--no-sync`	Generate your personal year-in-review coding-agent summary

Before anything is generated, prompts are redacted client-side: secrets matching AWS/OpenAI/GitHub/Slack key patterns, UUIDs, and long hex strings are stripped, each prompt is truncated to 500 characters, and the whole corpus is capped at 2,000 prompts / 500K characters. Only text you actually typed is used — never assistant responses or tool input/output. The backend re-redacts as a second layer of defense.

Troubleshooting

Symptom	Cause	Fix
`entelligence auth status` shows not authenticated	No API key in config or env	Run `entelligence auth login`, or set `ENTELLIGENCE_API_KEY`
"No changes detected" on `review`	No uncommitted or committed diff found	Check `git status`; use `--committed-only --base-branch <branch>` if your changes are already committed
Base URL not switching to local dev	`ENTELLIGENCE_BASE_URL` not set / stale `.env`	`export ENTELLIGENCE_BASE_URL=http://localhost:8000`, or check `.env` in an editable install
Command hangs or times out	Network issue reaching the API	Re-run with `--debug` (on `review`) to see the raw request, and check `~/.entelligence/cli-debug.log`
Config permission errors	Corrupted `~/.entelligence/config.json`	`rm ~/.entelligence/config.json && entelligence auth login`
Changes to an editable/dev install aren't reflected	Stale install	`uv tool uninstall entelligence-cli && make setup` (from the CLI repo)

💡

Every command supports --help for the exact, current flag list on your installed version — e.g. entelligence review --help.

Overview Overview