
Caleb RhodesPaperclip can use Groniz for social publishing through a skill, MCP, or CLI. To set up the skill,...
Paperclip can use Groniz for social publishing through a skill, MCP, or CLI. To set up the skill, check three places before asking an agent to schedule anything: the company skill library, the agent's assigned skills, and the runtime that will execute the commands.
After importing the skill, check the environment that will run it. The executing environment still needs the Groniz CLI and working credentials. Once those are available, the agent can discover a destination and prepare a post for review.
The steps follow the documented setup paths. We have not tested a complete delivery inside Paperclip. The client-by-client setup guide gives the broader context.
Record a result for each row in this worksheet. For example, an imported skill still needs to reach the agent that will use it.
| Setup step | Question | Evidence to record | If incomplete |
|---|---|---|---|
| Company library | Is the Groniz skill imported into the intended company? | Company ID and imported skill entry | Resolve the company import |
| Agent attachment | Is the intended agent assigned that skill? | Agent ID and desired skill set | Update the agent's assigned skills |
| Runtime | Can that agent's execution environment run and authenticate Groniz? | Executable check and sanitized identity result | Fix runtime availability or credentials |
Paperclip's CLI documentation distinguishes company library imports, agent attachment, and adapter runtime synchronization. Treat them as separate layers when investigating why an agent cannot use an imported skill. Paperclip CLI documentation
This worksheet also identifies who can fix a problem. A content reviewer may be able to confirm the destination but have no control over the agent runtime. Record which check failed so the next operator knows where to begin.
Check the installed CLI's help before applying these documented commands. Confirm the intended company and agent identifiers, and review the agent's existing skill assignments:
paperclipai skills import groniz/groniz-cli --company-id YOUR_COMPANY
paperclipai skills agent sync YOUR_AGENT --desired-skills groniz
Check the full list passed to --desired-skills. Preserve any other assignments the agent needs when constructing the desired set for your installation; do not unintentionally replace its whole skill list with Groniz alone. Use the installed command help to verify the supported syntax. Paperclip CLI documentation
Record the company and agent IDs in your setup notes. Avoid using "the marketing agent" as the only identifier when several agents or companies have similar names.
Then verify what reached the actual runtime. A successful import or assignment does not prove the execution process received the required files and environment.
Run these checks in the environment that will execute the publishing task:
command -v groniz
groniz whoami
groniz integrations:list
If the executable is missing, the runtime setup is incomplete. If whoami fails, check the runtime credentials through your approved mechanism. Keep tokens out of command transcripts and shared setup notes. If the account list is available, choose the intended integration by its returned ID and retrieve its requirements:
groniz integrations:settings INTEGRATION_ID
Replace INTEGRATION_ID with the selected integration ID. The returned schema is the source of truth for required settings, limits, and provider-specific requirements. Groniz handles OAuth, platform formatting, and delivery, but capabilities vary by provider.
You can run these checks before choosing a post to publish. Confirm that the runtime has an authenticated CLI, then ask the agent to prepare a delivery.
Groniz's documented Paperclip integration registers write tools as destructive and quarantines them until an operator enables them in the catalog. Follow that operator step for the tools you intend to use. Catalog availability does not approve the content of a particular post.
Before execution, ask the agent to show a complete review record:
Agent and company: [exact identifiers]
Destination: [integration ID and readable account label]
Final body: [all text and links]
Media: [uploaded Groniz .path values, or none]
Provider settings: [values checked against the current schema]
Schedule: [ISO timestamp and intended timezone]
Authorized action: [one specific create or schedule operation]
This is a review template, not a built-in Paperclip approval object. Your team decides how approval is recorded. The executor needs to know exactly which material and action the reviewer approved. The human approval guide explains how to handle later edits.
If the post uses media, upload it first and carry forward the returned .path. A file mentioned in a planning document may not be accessible to the runtime, and a local path is not the uploaded media reference.
After an authorized submission, save the returned post ID with the review record. Describe an accepted scheduled post as queued until publication is confirmed.
If the agent ends with a timeout or incomplete response, inspect the remote queue before issuing another create request. A replacement agent should receive the existing attempt record so it can reconcile the result rather than repeat the write.
Begin by connecting the intended social account in Groniz, then check the company, skill assignment, and runtime for the Paperclip agent that will execute the task.