CLI commands - ai-cmo.dev

Each section below covers one subcommand, its flags, and what it prints.

`geo init`

Initialize project directories and database. Flags: None Creates db/, raw/, cache/, reports/ directories under the data root (default ./data). Initializes the SQLite schema at data/db/geo.sqlite3. Loads all brand YAML files, engines, and presets, printing a summary. Requires four environment variables: OPENROUTER_API_KEY, GOOGLE_WORKSPACE_CLI_CLIENT_ID, GOOGLE_WORKSPACE_CLI_CLIENT_SECRET, GOOGLE_ANALYTICS_REFRESH_TOKEN.

`geo status`

Show database and configuration status. Flags: None Prints the database path, file size, table row counts, last run per brand, and pending extraction count.

`geo run`

Run brand-visibility prompts.

Flag	Type	Default	Description
`--operation-id`	string	—	Server-issued metered operation id
`--brand`	string	—	Brand key from `brands/<key>.yaml`
`--all-brands`	flag	false	Run every configured brand
`--preset`	string	—	Preset name (default: `standard`)
`--engines`	string	—	Comma-separated engine keys
`--smoke`	flag	false	Force the smoke preset
`--resume`	string	—	Resume an existing run by RUN_ID
`--dry-run`	flag	false	Print the matrix and cost without network calls

Prints the prompt/engine matrix, cost summary, and run ID on completion.

`geo research run`

Run evidence-grounded market and ICP research (five-stage frozen DAG).

Flag	Type	Default	Description
`--operation-id`	string	required	Server-issued metered operation id

Prints the research run ID and artifact count.

`geo extract`

Extract mentions, citations, and sentiment from responses.

Flag	Type	Default	Description
`--brand`	string	—	Brand key from `brands/<key>.yaml`
`--run`	string	—	Process only responses from this run
`--reextract`	flag	false	Reprocess already-extracted responses with new schema version
`--schema-version`	string	`v1`	Schema version tag

Prints per-response mention/citation counts, total cost, and error count.

`geo prompts`

Manage generated prompt sets. Requires a subcommand.

`geo prompts seed`

Build GSC and brain seed corpus.

Flag	Type	Default	Description
`--brand`	string	required	Brand key

Prints the number of seeds written.

`geo prompts generate`

Generate candidate prompts.

Flag	Type	Default	Description
`--brand`	string	optional	Brand key (inferred from metering when `--operation-id` is set)
`--operation-id`	string	—	Server-issued metered operation id
`--dry-run`	flag	false	Print planned cells without LLM calls

When --operation-id is set, performs metered, token-authorized generation. When omitted (and --dry-run is set), uses the ambient brand config.

`geo prompts curate`

Fill the active prompt quotas.

Flag	Type	Default	Description
`--brand`	string	required	Brand key

Prints the number of activated prompts and coverage matrix.

`geo prompts discriminate`

Run the adversarial critic loop.

Flag	Type	Default	Description
`--brand`	string	required	Brand key

Prints approval rate, round count, and paraphrase count.

`geo prompts relabel`

Replace lexical topic labels with domain topics.

Flag	Type	Default	Description
`--brand`	string	required	Brand key
`--dry-run`	flag	false	Print the old-to-new mapping without database writes

Interactive: prompts for confirmation before applying.

`geo prompts status`

Show the prompt coverage matrix.

Flag	Type	Default	Description
`--brand`	string	required	Brand key

`geo prompts add`

Add a user-authored prompt.

Flag	Type	Default	Description
`--brand`	string	required	Brand key
`--text`	string	required	Prompt text
`--intent`	string	required	Intent (must be in brand quotas)
`--topic`	string	required	Topic label
`--lang`	string	required	Language code
`--importance`	float	—	Importance weight

Prints the new prompt ID.

`geo prompts edit`

Edit a prompt’s fields.

Flag	Type	Default	Description
`--brand`	string	required	Brand key
`--id`	string	required	Prompt ID
`--text`	string	—	New text
`--intent`	string	—	New intent
`--topic`	string	—	New topic
`--lang`	string	—	New language
`--importance`	float	—	New importance

Prints the updated or archived-and-recreated prompt ID.

`geo prompts archive`

Deactivate a prompt.

Flag	Type	Default	Description
`--brand`	string	required	Brand key
`--id`	string	required	Prompt ID

`geo prompts restore`

Reactivate a prompt.

Flag	Type	Default	Description
`--brand`	string	required	Brand key
`--id`	string	required	Prompt ID

`geo recs`

List, inspect, and update recommendations. Requires a subcommand.

`geo recs generate`

Create a metered, idempotent recommendation batch.

Flag	Type	Default	Description
`--operation-id`	string	required	Server-issued metered operation id
`--idempotency-key`	string	required	Unique idempotency key

Prints the generated recommendations as JSON.

`geo recs list`

List recommendations for a brand.

Flag	Type	Default	Description
`--brand`	string	required	Brand key
`--status`	string	`proposed`	Recommendation status to filter
`--json`	flag	false	Print a JSON array

Prints a formatted table or JSON array.

`geo recs show`

Show a complete recommendation.

Flag	Type	Default	Description
`rec_id`	positional	required	Recommendation ID
`--json`	flag	false	Print JSON
`--include-transitions`	flag	false	Include transition history

`geo recs draft`

Generate a draft comment for a Reddit recommendation.

Flag	Type	Default	Description
`--brand`	string	required	Brand key
`--id`	string	required	Recommendation ID

Prints the draft as JSON.

`geo recs transition`

Execute a CAS-guarded lifecycle transition.

Flag	Type	Default	Description
`--brand`	string	required	Brand key
`--id`	string	required	Recommendation ID
`--action`	string	required	Action (one of the canonical actions)
`--expected-status`	string	required	Expected current status
`--expected-version`	int	required	Expected version
`--idempotency-key`	string	required	Unique key
`--actor-kind`	string	`session`	Actor kind
`--actor-id`	string	`cli-user`	Actor identifier
`--note`	string	—	Transition note
`--review-evidence`	string	—	JSON string of review evidence
`--visibility-snapshot`	string	—	JSON string of visibility snapshot
`--compliance-approved`	flag	false	Mark compliance reviewed
`--draft-id`	string	—	Draft identifier
`--metadata`	string	—	JSON object attached to the audit row
`--json`	flag	false	Print JSON result

`geo recs mark`

[DEPRECATED] Change recommendation status (maps to canonical transitions).

Flag	Type	Default	Description
`--brand`	string	required	Workspace/brand key
`rec_id`	positional	required	Recommendation ID
`status`	positional	required	One of: `accepted`, `shipped`, `verified`, `rejected`, `reopened`
`--note`	string	—	Append an artifact or status note
`--expected-status`	string	required	Expected current status
`--expected-version`	int	required	Expected version
`--idempotency-key`	string	required	Unique key
`--actor-kind`	string	`session`	Actor kind
`--actor-id`	string	`cli-user`	Actor identifier
`--review-evidence`	string	—	JSON string of review evidence
`--visibility-snapshot`	string	—	JSON string of visibility snapshot
`--draft-id`	string	—	Draft identifier
`--force-compliance-reviewed`	flag	false	Confirm compliance review before shipping

`geo seo`

Run deterministic SEO scoring, insights, and audits. Requires a subcommand.

`geo seo score`

Compute SEO score.

Flag	Type	Default	Description
`--operation-id`	string	required	Server-issued metered operation id
`--date`	string	—	Score/window end date (YYYY-MM-DD)

`geo seo insights`

Build SEO insights.

Flag	Type	Default	Description
`--operation-id`	string	required	Server-issued metered operation id
`--date`	string	—	Score/window end date (YYYY-MM-DD)

`geo seo audit`

Run a D4 SEO audit.

Flag	Type	Default	Description
`--operation-id`	string	required	Server-issued metered operation id

`geo metrics`

Compute daily visibility metrics.

Flag	Type	Default	Description
`--recompute`	flag	false	Delete and rewrite selected `metrics_daily` rows
`--since`	string	—	Only compute run dates on or after YYYY-MM-DD

Prints row count, date slices, and calculation version.

`geo report`

Render Markdown and JSON reports.

Flag	Type	Default	Description
`--brand`	string	required	Brand key from `brands/<key>.yaml`
`--date`	string	—	Run date (YYYY-MM-DD, defaults to latest)

Prints the Markdown and JSON output file paths.

Generate evidence-grounded GEO recommendations.

Flag	Type	Default	Description
`--brand`	string	required	Brand key from `brands/<key>.yaml`
`--window`	int	28	Trailing analysis window in days
`--dry-run`	flag	false	Print gap analysis without LLM calls or writes

geo recommend without --dry-run always fails by design. Recommendation creation has moved to the token-authorized geo recs generate command. Run geo recs generate --operation-id ... --idempotency-key ... instead.

When --dry-run is set, prints a gap analysis table with no side effects.

`geo daily`

Run the complete per-brand daily pipeline.

Flag	Type	Default	Description
`--preset`	string	`standard`	Preset name
`--brands`	string	`didit,hi-doctor,sante`	Comma-separated brand keys
`--dry-run`	flag	false	Print the per-brand plan without writes or calls

`geo validate`

Run GEO validation gates.

Flag	Type	Default	Description
`--brand`	string	—	Validate only one brand key
`--gate`	string	`all`	Gate to run: `all`, `prompts`, `extraction`, `metrics`, `e2e`
`--judge`	flag	false	Run the live third-family judge prompt sub-gate
`--loop`	flag	false	Retry only failing gates up to three times

`geo costs`

Not implemented. This command is a stub that prints 'costs' not implemented and exits with code 2.

​geo init

​geo status

​geo run

​geo research run

​geo extract

​geo prompts

​geo prompts seed

​geo prompts generate

​geo prompts curate

​geo prompts discriminate

​geo prompts relabel

​geo prompts status

​geo prompts add

​geo prompts edit

​geo prompts archive

​geo prompts restore

​geo recs

​geo recs generate

​geo recs list

​geo recs show

​geo recs draft

​geo recs transition

​geo recs mark

​geo seo

​geo seo score

​geo seo insights

​geo seo audit

​geo metrics

​geo report

​geo recommend

​geo daily

​geo validate

​geo costs

`geo init`

`geo status`

`geo run`

`geo research run`

`geo extract`

`geo prompts`

`geo prompts seed`

`geo prompts generate`

`geo prompts curate`

`geo prompts discriminate`

`geo prompts relabel`

`geo prompts status`

`geo prompts add`

`geo prompts edit`

`geo prompts archive`

`geo prompts restore`

`geo recs`

`geo recs generate`

`geo recs list`

`geo recs show`

`geo recs draft`

`geo recs transition`

`geo recs mark`

`geo seo`

`geo seo score`

`geo seo insights`

`geo seo audit`

`geo metrics`

`geo report`

`geo recommend`

`geo daily`

`geo validate`

`geo costs`