bolta-skills

Activate a paused job after preview and voice validation - the explicit trust moment where user says "yes, start posting"

Modify an existing agent's configuration including persona, model tier, enabled skills, and job settings

Create and onboard a new AI agent teammate from marketplace presets with conversational discovery and preview generation.

Store and retrieve information across job runs - how agents learn and improve over time

Handle @mention interactions where users ask agents for quick feedback on posts and drafts

Bolta Skills Registry - canonical index and orchestration layer for all Bolta skills, organized by plane

Export audit/activity events for a workspace (agent + human actions) for debugging and compliance.

Explain workspace policy and the caller's effective capabilities (what actions are allowed) in plain language.

View current quota usage and limits for a workspace including post count, API requests, and reset timers

Create an agent identity (service account) inside a workspace and issue an API key with least privilege.

Rotate or revoke an existing API key for a human or agent principal, issuing a replacement key.

View and update workspace-level configuration settings including Safe Mode, autonomy mode, and quotas

Add a review comment to a draft without approving or rejecting - used for suggestions, questions, or notes

Approve a draft post, moving it from Inbox to Approved status, ready for scheduling or publishing

V2 Reviewer agent skill - Cluster inbox items, flag risks, propose edits, optionally create improved variants. Makes the inbox manageable.

Reject a draft post with specific feedback, sending it back for revision to enforce brand standards

Approve inbox posts and route to Approved→Scheduled based on workspace governance. Handles bulk approvals with flexible scheduling.

V2 - Generate a human-readable review summary (daily/weekly digest) of inbox activity, agent performance, and pending actions.

Evolve a voice profile by learning from new content samples. V2-ready - continuous voice refinement.

Documentation of V2 execution engine - how jobs fire and agents reason through tasks (NOT a callable tool)

Retrieve audience demographics, behavior patterns, and engagement trends for data-driven content strategy

Analyze historical performance to recommend optimal posting schedule based on when YOUR specific audience is most engaged

Retrieve performance metrics for published posts including likes, comments, shares, views, and engagement rate

V2 Job execution skill with direct scheduling - Agent generates N posts and schedules them ONLY if Safe Mode OFF AND agent is Editor/Admin. Otherwise routes to Inbox.

V2 Job execution skill - Agent generates N posts per run, routes to Inbox for review. Called by Celery job runner. Replaces RecurringTemplate logic.

Generate a 7-day content calendar draft using voice profile + optional templates. V2 - Agent-aware, routes via workspace Safe Mode governance.

Fetch comments on posts to monitor conversations and identify response opportunities or policy violations

Fetch recent mentions, replies, and comments directed at the account for community engagement monitoring

Draft a reply to a mention, comment, or DM while maintaining brand voice and community presence

Create a comprehensive Voice Profile for a workspace. V2-ready - used by all agent types before content generation.

Create a draft post using workspace voice profile - routes to Inbox for review before publishing

Get social account details including platform, handle, status, and token validity

Get business context including products, audience, values, and brand DNA for informed content creation

Retrieve voice profile details including tone, style, vocabulary, dos/donts to understand brand voice before creating content

Get the last N published posts for an account to avoid repetition and maintain content variety

Create a recurring content loop from a template using a voice profile, then send all generated posts into Inbox Review as a bundle.