Bolta.skills.index

Metadata

{
  "name": "bolta.skills.index",
  "version": "2.0.0",
  "publisher": "bolta.ai",
  "verified": true,
  "sourceRepository": "https://github.com/boltaai/bolta-skills",
  "requiredEnvironmentVariables": [
    {
      "name": "BOLTA_API_KEY",
      "required": true,
      "sensitive": true,
      "description": "Bolta API key (obtain at bolta.ai/register)",
      "format": "sk_live_[64 characters]",
      "scope": "workspace"
    },
    {
      "name": "BOLTA_WORKSPACE_ID",
      "required": true,
      "sensitive": false,
      "description": "Workspace UUID for API operations",
      "format": "UUID"
    },
    {
      "name": "BOLTA_AGENT_ID",
      "required": false,
      "sensitive": false,
      "description": "Agent principal UUID (for audit logging)",
      "format": "UUID"
    }
  ],
  "trustedDomains": [
    "platty.boltathread.com",
    "bolta.ai",
    "mcp.bolta.ai"
  ],
  "permissions": [
    "network:https:platty.boltathread.com",
    "network:https:bolta.ai",
    "network:https:mcp.bolta.ai"
  ],
  "relatedRepositories": [
    {
      "name": "bolta-skills",
      "description": "Official Bolta Skills Pack",
      "url": "https://github.com/boltaai/bolta-skills"
    },
    {
      "name": "boltaclaw-self-hosted",
      "description": "Self-hosted Bolta agent runtime (BoltaClaw)",
      "url": "https://github.com/boltaai/boltaclaw-self-hosted"
    }
  ],
  "mcpEndpoint": "https://mcp.bolta.ai/mcp",
  "apiDocumentation": "https://bolta.ai/docs/api"
}

Security Notice

This skill requires sensitive API credentials. Read this section carefully before installing.

Required Credentials

BOLTA_API_KEY (REQUIRED, SENSITIVE)

• Format: bolta_sk_ followed by alphanumeric characters
• Obtain at: https://bolta.ai/register (agent keys) or bolta.ai/settings (regular keys)
• Scoping: Each key is scoped to a SINGLE workspace only
• Key Types: Bolta issues two distinct key types (see API Key Types below)
• Rotation: Rotate every 90 days using bolta.team.rotate_key skill
• Storage: NEVER commit to git - use environment variables or secret managers only

BOLTA_WORKSPACE_ID (REQUIRED)

• Format: UUID (e.g., 550e8400-e29b-41d4-a716-446655440000)
• Source: Provided during agent registration at bolta.ai/register
• Purpose: Identifies which workspace the API key is authorized for

BOLTA_AGENT_ID (OPTIONAL — required for Agent API keys)

• Format: UUID
• Purpose: Links API activity to specific agent principal for audit logs
• Benefit: Enables traceability and compliance reporting
• Note: Automatically set when using an Agent API key; must be provided manually with Regular API keys if audit attribution is desired

API Key Types

Bolta has two distinct API key types with different permission models. Skills must handle both correctly.

Regular API Keys (key_type: "regular")

Created in Settings → API Keys by workspace members. Use granular permission scopes — each key is issued with an explicit list of allowed scopes.

Available scopes:

Permission groups (shortcuts):

• READ_ONLY — posts:read, accounts:read, voice:read
• CONTENT_CREATOR — READ_ONLY + posts:write, ai:generate
• FULL_ACCESS — All scopes

When to use: Integrations, CI/CD pipelines, custom scripts, third-party tools that need specific scopes.

Agent API Keys (key_type: "agent")

Created in Settings → AI Agents → Add Agent by workspace admins. Use role-based permissions — the key inherits all scopes from the assigned agent role, plus agent-specific scopes not available to regular keys.

Agent roles and their scopes:

Agent-only scopes (not available on regular keys):

Additional agent key behaviors:

• Safe Mode is always enforced for creator role agents
• Review required — creator agents cannot publish directly; content goes through approval
• Agent keys carry an agent_id that is automatically used for audit attribution
• Agent keys link to an AgentPrincipal record for identity tracking

Skill Plane Access by Role

Agent roles determine which skill planes are accessible. Skills should check the caller's role before executing plane-restricted operations.

A viewer can access Review and Control planes. A creator can access all planes. editor and admin can access all planes with additional capabilities within each.

How Skills Should Handle Key Types

1. Check key_type on the API response — the server returns key_type: "regular" | "agent" on authenticated responses
2. Regular keys: Validate required scopes exist in the key's permissions array before executing
3. Agent keys: The role determines capabilities — check the role rather than individual scopes
4. Agent-only operations (review workflows, audit export, team management) — reject if key_type is "regular" since these scopes are not available
5. Audit attribution: If key_type is "agent", the agent_id is automatically attached; for regular keys, pass BOLTA_AGENT_ID env var if attribution is needed

Trusted Network Endpoints

This skill makes HTTPS requests to:

• https://platty.boltathread.com - Bolta API server
• https://bolta.ai - Main application and agent registration portal
• https://mcp.bolta.ai - MCP protocol endpoint (for Claude Desktop / Claude Code)

No other domains are contacted. All requests are authenticated with your API key.

Related Projects & Resources

• Bolta Skills Pack — Official skills repository
  • Source: https://github.com/boltaai/bolta-skills
• BoltaClaw — Self-hosted Bolta agent runtime
  • Source: https://github.com/boltaai/boltaclaw-self-hosted
• Bolta MCP Server — Remote MCP endpoint for Claude Desktop / Claude Code integration
  • Endpoint: https://mcp.bolta.ai/mcp
• Bolta API Documentation — Full API reference
  • Docs: https://bolta.ai/docs/api

Pre-Installation Checklist

Before installing this skill, you MUST:

• Verify the source repository: https://github.com/boltaai/bolta-skills
• Review the SKILL.md and confirm version matches metadata (currently 2.0.1)
• Obtain a LEAST-PRIVILEGE API key from https://bolta.ai/register
• Store API key in environment variables (NEVER hardcode or commit)
• Verify you trust the domains: platty.boltathread.com, bolta.ai, and mcp.bolta.ai
• Test in a disposable/test workspace first (recommended)
• Confirm your API key is scoped ONLY to the intended workspace

Security Best Practices

1. Credential Management

   • Use environment variables: export BOLTA_API_KEY="sk_live_..."
   • Or use secret managers: AWS Secrets Manager, 1Password, etc.
   • NEVER paste API keys in chat, logs, or public places

2. Key Rotation

   • Rotate keys every 90 days minimum
   • Use bolta.team.rotate_key skill for zero-downtime rotation
   • Revoke compromised keys immediately at bolta.ai/settings

3. Permission Scoping

   • Grant ONLY required permissions (e.g., posts:write, voice:read)
   • Avoid workspace:admin unless absolutely necessary
   • Review permissions quarterly

4. Monitoring

   • Review audit logs weekly via bolta.audit.export_activity
   • Monitor quota usage via bolta.quota.status
   • Set up alerts for unusual API activity

5. Workspace Isolation

   • One API key per workspace (NEVER share keys across workspaces)
   • Use separate keys for dev/staging/production environments
   • Revoke keys when decommissioning workspaces

Purpose

The canonical registry and orchestration layer for all Bolta skills.

This skill serves as the single source of truth for skill discovery, installation recommendations, and workspace-aware capability bootstrapping. It does not execute content operations directly — instead, it provides intelligent routing to the appropriate skills based on:

• Workspace policy (Safe Mode, autonomy mode, quotas)
• Principal identity (user role, agent permissions)
• Operational context (what you're trying to accomplish)

Key Responsibilities:

1. Discovery - Index all available skills with metadata
2. Recommendation - Suggest install sets based on workspace policy and role
3. Orchestration - Guide multi-skill workflows
4. Compatibility - Enforce skill compatibility with workspace settings
5. Bootstrapping - Help new workspaces get started quickly

When to Use:

• Setting up a new workspace ("What skills should I install?")
• Discovering available capabilities ("What can Bolta do?")
• Troubleshooting skill compatibility ("Why can't I use this skill?")
• Planning multi-step workflows ("Which skills do I need?")

Data Access: This skill accesses:

• Workspace configuration (policy, quotas, autonomy mode)
• Voice profile metadata (names, IDs, not full content)
• Post counts and quota usage
• Agent principal types

This skill does NOT access:

• Post content or scheduled posts
• Social media credentials
• User passwords or authentication tokens
• Files or media uploads

Source & Verification

https://github.com/boltaai/bolta-skills

Bolta Ecosystem: How It All Fits Together

Bolta is composed of four interconnected pieces. Understanding how they relate helps you choose the right setup path.

┌─────────────────────────────────────────────────────────────┐
│                      bolta.ai                               │
│            Workspace management, dashboards,                │
│            account connections, review UI                    │
│                                                             │
│  API Docs: https://bolta.ai/docs/api                       │
└──────────────────────┬──────────────────────────────────────┘
                       │ REST API
                       ▼
┌──────────────────────────────────────────────────────────────┐
│                  MCP Server (mcp.bolta.ai/mcp)               │
│     Bridges Claude Desktop / Claude Code to the Bolta API    │
│     Exposes workspace tools: create posts, manage agents,    │
│     approve content, run analytics — all via MCP protocol    │
└──────────────────────┬───────────────────────────────────────┘
                       │ uses
                       ▼
┌──────────────────────────────────────────────────────────────┐
│            Bolta Skills Pack (this repo)                      │
│     github.com/boltaai/bolta-skills                          │
│                                                              │
│     Structured prompts, workflows, and orchestration logic   │
│     that teach AI agents HOW to use the API effectively:     │
│     voice creation, content planning, review flows,          │
│     automation, analytics, engagement, governance            │
└──────────────────────────────────────────────────────────────┘

┌──────────────────────────────────────────────────────────────┐
│            BoltaClaw (turnkey self-hosted runtime)            │
│     github.com/boltaai/boltaclaw-self-hosted                 │
│                                                              │
│     Everything above — pre-configured and ready to run.      │
│     Ships with the skills pack, MCP integration, agent       │
│     runtime, cron scheduler, and CLI out of the box.         │
│     Clone, set your API key, and go.                         │
└──────────────────────────────────────────────────────────────┘

How they work together

1. The API (bolta.ai/docs/api) is the foundation — every operation (creating posts, managing agents, approving content) flows through it. You can call it directly with curl or any HTTP client.

2. The MCP Server (mcp.bolta.ai/mcp) wraps the API into the MCP protocol so Claude Desktop and Claude Code can call Bolta tools natively. Point your MCP config at the endpoint and your AI assistant gains full workspace access.

3. The Skills Pack (github.com/boltaai/bolta-skills) provides the intelligence layer — structured prompts and multi-step workflows that teach agents how to use the API well. Skills handle things like voice-aware drafting, week planning, review triage, and cron automation that go beyond raw API calls.

4. BoltaClaw (github.com/boltaai/boltaclaw-self-hosted) is the turnkey solution. It bundles the skills pack, MCP connectivity, agent runtime, job scheduler, and a CLI into a single self-hosted package. The CLI lets you manage posts, agents, jobs, and workspace settings directly from your terminal — no browser required. If you want to skip manual setup and get a fully operational Bolta agent running on your own infrastructure, BoltaClaw is the fastest path — clone, configure your API key, and you're live.

Which path is right for you?

Getting Started: API Setup

Before using Bolta skills, you need an API key. Choose the path that matches your use case.

Path A: Agent API Key (Recommended for AI Agents)

Use this if you're setting up an AI agent (Claude, BoltaClaw, custom agent) to operate as a workspace member.

Step 1: Create Agent Principal

Go to Settings → AI Agents → Add Agent in your Bolta workspace.

Configure:

• Agent Name — Human-readable name for audit logs (e.g., "Claude Content Agent")
• Agent Role — Determines permissions and skill plane access:

Recommendation: Start with creator role. Safe Mode is enforced and all content requires human review.

Step 2: Copy Your Agent API Key

After creation, you'll see the full key once:

API Key: bolta_sk_00000000000000000000000000000000
Workspace ID: 550e8400-e29b-41d4-a716-446655440000
Agent ID: 660e8400-e29b-41d4-a716-446655440001

The agent ID is automatically linked — no need to set BOLTA_AGENT_ID separately.

Path B: Regular API Key (For Integrations & Scripts)

Use this for CI/CD pipelines, custom integrations, or scripts that need specific scopes.

Step 1: Create API Key

Go to Settings → API Keys → Create New Key.

Select only the scopes you need (least privilege):

• Content creation: posts:write, voice:read
• Full content: posts:read, posts:write, voice:read, ai:generate
• Read-only: posts:read, accounts:read, voice:read

Note: Regular keys cannot access agent-only scopes (review:submit, review:approve, audit:export, team:manage, team:manage_keys, workspace:read). If you need those, use an Agent API key instead.

Step 2: Copy Your API Key

API Key: bolta_sk_00000000000000000000000000000000

IMPORTANT: The full key is shown only once and cannot be recovered (only regenerated via bolta.team.rotate_key).

Configure Your Environment (Both Key Types)

Set Required Environment Variables:

# Required: Your Bolta API key (agent or regular)
export BOLTA_API_KEY="bolta_sk_your_actual_key_here"

# Required: Your workspace UUID
export BOLTA_WORKSPACE_ID="550e8400-e29b-41d4-a716-446655440000"

# Optional: Agent principal UUID (automatic for agent keys, manual for regular keys)
export BOLTA_AGENT_ID="660e8400-e29b-41d4-a716-446655440001"

For Claude Desktop / Claude Code (MCP):

Connect via the remote MCP endpoint at https://mcp.bolta.ai/mcp:

{
  "mcpServers": {
    "bolta": {
      "type": "url",
      "url": "https://mcp.bolta.ai/mcp",
      "headers": {
        "Authorization": "Bearer bolta_sk_your_actual_key_here",
        "X-Workspace-ID": "550e8400-e29b-41d4-a716-446655440000"
      }
    }
  }
}

For self-hosted deployments, see BoltaClaw for running your own agent runtime.

For Direct API Calls:

See the full API documentation for all available endpoints.

curl https://platty.boltathread.com/v1/posts \
  -H "Authorization: Bearer ${BOLTA_API_KEY}" \
  -H "X-Workspace-ID: ${BOLTA_WORKSPACE_ID}" \
  -H "Content-Type: application/json" \
  -d '{ "prompt": "Create a post about productivity" }'

Security Reminder:

• Never hardcode API keys in your code
• Use .env files locally (add .env to .gitignore)
• Use secret managers in production (AWS Secrets Manager, Vercel Secrets, etc.)
• Rotate keys every 90 days via bolta.team.rotate_key

Verify Setup

Test your configuration:

curl https://platty.boltathread.com/v1/workspaces/${BOLTA_WORKSPACE_ID} \
  -H "Authorization: Bearer ${BOLTA_API_KEY}"

# Expected response:
# {
#   "id": "550e8400-...",
#   "name": "My Workspace",
#   "safe_mode": true,
#   "autonomy_mode": "managed",
#   "max_posts_per_day": 100,
#   "key_type": "agent",        ← indicates which key type authenticated
#   "agent_role": "creator"     ← present only for agent keys
# }

Troubleshooting Setup

Error: "Invalid API Key"

• Verify key matches exactly (no extra spaces)
• Check if key was rotated — get new key at bolta.ai/settings
• Ensure you're using the correct workspace key
• Confirm key prefix starts with bolta_sk_

Error: "Workspace Not Found"

• Verify workspace_id matches your registration
• Confirm you have access to this workspace (visit bolta.ai/workspaces)
• Check if workspace was deleted

Error: "Permission Denied"

• Regular keys: Check granted scopes at Settings → API Keys
  • Content creation: Need posts:write minimum
  • Voice access: Need voice:read or voice:write
• Agent keys: Check agent role at Settings → AI Agents
  • Content creation: Need creator role or higher
  • Approvals: Need editor role or higher
  • Agent management: Need admin role
• Agent-only scopes (review:submit, audit:export, etc.) are not available on regular keys — switch to an Agent API key if you need them

Error: "Insufficient Role" (Agent keys only)

• The agent's role doesn't have access to the requested skill plane
• viewer cannot access Init, Content, or Automation planes
• Upgrade the agent role at Settings → AI Agents, or use a different agent

Installation & First Run

Skill Pack Installation

Option 1: Install Full Skill Pack (Recommended)

# Clone the complete Bolta skills repository
git clone https://github.com/boltaai/bolta-skills.git

# Or download the latest release
curl -L https://github.com/boltaai/bolta-skills/archive/refs/heads/main.zip -o bolta-skills.zip
unzip bolta-skills.zip

What You Get:

bolta-skills/
├── README.md                    # START HERE - Complete getting started guide
├── skills/
│   ├── bolta.skills.index/      # You're here (registry)
│   ├── voice-plane/
│   │   ├── bolta.voice.bootstrap/
│   │   └── bolta.voice.learn_from_samples/
│   ├── content-plane/
│   │   ├── bolta.draft.post/
│   │   ├── bolta.draft_post/
│   │   ├── bolta.get_account_info/
│   │   ├── bolta.get_business_context/
│   │   ├── bolta.get_voice_profile/
│   │   ├── bolta.list_recent_posts/
│   │   ├── bolta.loop.from_template/
│   │   └── bolta.week.plan/
│   ├── review-plane/
│   │   ├── bolta.add_comment/
│   │   ├── bolta.approve_post/
│   │   ├── bolta.reject_post/
│   │   ├── bolta.inbox.triage/
│   │   ├── bolta.review.approve_and_route/
│   │   └── bolta.review.digest/
│   ├── automation-plane/
│   │   ├── bolta.cron.generate_to_review/
│   │   └── bolta.cron.generate_and_schedule/
│   ├── agent-plane/
│   │   ├── bolta.agent.activate_job/
│   │   ├── bolta.agent.configure/
│   │   ├── bolta.agent.hire/
│   │   ├── bolta.agent.memory/
│   │   ├── bolta.agent.mention/
│   │   └── bolta.job.execute/
│   ├── analytics-plane/
│   │   ├── bolta.get_audience_insights/
│   │   ├── bolta.get_best_posting_times/
│   │   └── bolta.get_post_metrics/
│   ├── engagement-plane/
│   │   ├── bolta.get_comments/
│   │   ├── bolta.get_mentions/
│   │   └── bolta.reply_to_mention/
│   └── control-plane/
│       ├── bolta.team.create_agent_teammate/
│       ├── bolta.team.rotate_key/
│       ├── bolta.policy.explain/
│       ├── bolta.audit.export_activity/
│       ├── bolta.quota.status/
│       └── bolta.workspace.config/
├── docs/
│   ├── getting-started.md
│   ├── autonomy-modes.md
│   ├── safe-mode.md
│   ├── automation-setup.md
│   └── multi-agent.md

Option 2: Install Individual Skills (Manual)

# Install voice bootstrap skill
curl -L https://raw.githubusercontent.com/boltaai/bolta-skills/main/skills/voice-plane/bolta.voice.bootstrap/SKILL.md \
  -o bolta.voice.bootstrap.md

# Install draft post skill
curl -L https://raw.githubusercontent.com/boltaai/bolta-skills/main/skills/content-plane/bolta.draft.post/SKILL.md \
  -o bolta.draft.post.md

First Run: Read the README

IMPORTANT: After installation, read the README for complete setup instructions.

What the README Covers:

1. Prerequisites and system requirements
2. Installation verification
3. Environment variable configuration
4. First skill execution (test workflow)
5. Troubleshooting common issues
6. Recommended skill installation order
7. Best practices for production use

Recommended First-Run Flow

Step 1: Read Documentation

# Must-read files in order:
# 1. README.md              - Complete getting started guide
# 2. docs/getting-started.md - Quickstart tutorial
# 3. docs/autonomy-modes.md  - Understand autonomy levels
# 4. docs/safe-mode.md       - Understand safety controls

Step 2: Verify Installation

# Check that all skills are present
ls -la skills/*/SKILL.md skills/*/*/SKILL.md

# Should see 36+ skills across 8 planes

Step 3: Configure Agent

export BOLTA_API_KEY="sk_live_..."
export BOLTA_WORKSPACE_ID="..."

# Test API connectivity
curl https://platty.boltathread.com/v1/workspaces/${BOLTA_WORKSPACE_ID} \
  -H "Authorization: Bearer ${BOLTA_API_KEY}"

Step 4: Install Recommended Skills

# The registry will recommend skills based on your:
# - Safe Mode setting
# - Autonomy mode
# - User role
# - Current quotas

Common First-Run Mistakes

1. Skipping the README - Installing skills without reading README
2. Missing Environment Variables - Running skills without BOLTA_API_KEY set
3. Installing Skills Out of Order - Running bolta.draft.post before creating voice profile
4. Not Understanding Autonomy Modes - Using autopilot mode without understanding routing
5. Hardcoding API Keys - Putting API keys directly in skill files instead of env vars

Post-Installation Checklist

• README.md has been read
• Environment variables set (BOLTA_API_KEY, BOLTA_WORKSPACE_ID)
• All 36+ skills present in skills/ directory
• docs/ directory contains markdown files
• API connectivity verified (test curl command works)
• MCP endpoint configured (if using Claude Desktop / Claude Code — see mcp.bolta.ai/mcp)
• Workspace policy reviewed (safe_mode, autonomy_mode)
• First skill executed successfully (test run)
• Autonomy mode documentation read (docs/autonomy-modes.md)
• Safe Mode documentation read (docs/safe-mode.md)

Next Steps After Setup

1. Create Voice Profile (if new workspace)

   • Run: bolta.voice.bootstrap
   • Establishes your brand voice

2. Test Content Creation

   • Run: bolta.draft.post or bolta.draft_post
   • Creates a test post in Draft status

3. Install Recommended Skills

   • Run: bolta.skills.index
   • Returns personalized skill recommendations

4. Configure Workspace

   • Review: Safe Mode (ON/OFF)
   • Review: Autonomy Mode (assisted/managed/autopilot)
   • Set: Daily quota limits

Architecture: The Eight Planes

Skills are organized into planes — logical groupings that separate concerns and enable modular capability composition. V2 expands from five planes to eight, adding Agent, Analytics, and Engagement planes.

Plane 0: Voice

Purpose: Brand voice creation, evolution, and validation

Voice is the foundation of all content operations. These skills help establish, refine, and maintain consistent brand voice across all generated content.

Core Principle: Voice should be learned from examples, validated against real content, and evolved over time.

Skills:

Typical Flow:

1. Bootstrap initial voice profile
2. Learn from sample content
3. Validate generated content
4. Evolve voice as brand matures

Plane 1: Content

Purpose: Content creation, planning, data retrieval, and scheduling

The execution layer for post creation. These skills transform ideas into scheduled social media posts, and provide workspace context for informed content decisions.

Core Principle: Content should be intentional, planned, and aligned with voice.

Skills:

Output: Draft or Scheduled posts (subject to Safe Mode routing)

Plane 2: Review

Purpose: Human-in-the-loop review, approval, and feedback workflows

Enables teams to review, approve, reject, and provide feedback on agent-generated content before publishing.

Core Principle: Autonomy with oversight — agents generate, humans decide.

Skills:

Typical Flow:

1. Agent creates posts -> Pending Approval
2. review.digest sends daily summary
3. Human reviews via inbox.triage
4. Approve/reject individual posts or bulk approve via approve_and_route

Plane 3: Automation

Purpose: Scheduled, recurring, and autonomous content generation

The autonomy layer. These skills enable hands-off content operations with guardrails.

Core Principle: Predictable automation with quota enforcement and safety nets.

Skills:

Safety Guardrails:

• Quota enforcement (max posts/day, max API requests/hour)
• Job run tracking (observability for all executions)
• Autonomy mode compatibility checks
• Safe Mode routing (autopilot incompatible with Safe Mode ON)

Plane 4: Agent

Purpose: AI agent lifecycle management — hiring, configuration, job activation, and memory

V2 introduces a dedicated agent plane for managing AI teammates that operate autonomously within your workspace.

Core Principle: Agents are first-class principals with defined roles, jobs, and memory.

Skills:

Typical Flow:

1. Hire an agent from presets (agent.hire)
2. Configure agent settings and jobs (agent.configure)
3. Activate the job to start autonomous work (agent.activate_job)
4. Agent builds memory over time (agent.memory)
5. Agents collaborate via mentions (agent.mention)

Plane 5: Analytics

Purpose: Performance measurement, audience insights, and posting optimization

Provides data-driven insights to inform content strategy and measure impact.

Core Principle: Measure what matters, optimize what works.

Skills:

Plane 6: Engagement

Purpose: Community interaction — monitoring mentions, replying, and managing conversations

Closes the feedback loop between publishing and audience interaction.

Core Principle: Respond authentically in brand voice, escalate what needs human judgment.

Skills:

Plane 7: Control

Purpose: Workspace governance, policy, security, and audit

The management layer for teams, permissions, security, and compliance.

Core Principle: Visibility and control for workspace administrators.

Skills:

Typical Use Cases:

• Onboarding new team members (human or agent)
• Investigating authorization failures
• Compliance reporting (SOC2, GDPR data exports)
• Quota monitoring and adjustment

Full Skill Index

Voice Plane Skills

bolta.voice.bootstrap

Path: skills/voice-plane/bolta.voice.bootstrap/SKILL.md Purpose: Interactive voice profile creation wizard Inputs: Brand name, industry, target audience, sample content Outputs: Complete VoiceProfile (tone, dos, don'ts, constraints) Permissions: voice:write Safe Mode: Compatible Roles: Viewer, Creator, Editor, Admin Typical Duration: 5-10 minutes (interactive)

bolta.voice.learn_from_samples

Path: skills/voice-plane/bolta.voice.learn_from_samples/SKILL.md Purpose: Extract voice patterns from existing content Inputs: URLs or text samples (3-10 examples) Outputs: Voice profile draft with auto-detected patterns Permissions: voice:write Safe Mode: Compatible Typical Duration: 2-3 minutes

bolta.voice.evolve (planned)

Purpose: Refine voice based on approved posts Inputs: Date range for approved posts Outputs: Updated VoiceProfile (version incremented) Permissions: voice:write, posts:read Safe Mode: Compatible

bolta.voice.validate (planned)

Purpose: Score content against voice profile Inputs: Post ID or content text Outputs: Compliance score (0-100), deviation report Permissions: voice:read, posts:read Safe Mode: Compatible

Content Plane Skills

bolta.draft.post

Path: skills/content-plane/bolta.draft.post/SKILL.md Purpose: Create a single post in Draft status (original interface) Inputs: Topic, platform(s), optional voice profile ID Outputs: Post ID (Draft status) Permissions: posts:write Safe Mode: Always routes to Draft Autonomy Mode: Respects assisted/managed routing Quota Impact: +1 to daily post count Typical Duration: 30-60 seconds

bolta.draft_post

Path: skills/content-plane/bolta.draft_post/SKILL.md Purpose: Create a draft post using workspace voice profile, routes to Inbox for review Inputs: content, platform, account_id, optional voice_profile_id Outputs: Post ID (Draft/Inbox status) Permissions: posts:write Safe Mode: Always routes to Inbox Roles: Viewer, Creator, Editor, Admin Safe Defaults: never_publish_directly, always_route_to_inbox

bolta.get_account_info

Path: skills/content-plane/bolta.get_account_info/SKILL.md Purpose: Retrieve connected social account details Inputs: account_id (optional, returns all if omitted) Outputs: Account metadata (platform, handle, status) Permissions: accounts:read Safe Mode: N/A (read-only)

bolta.get_business_context

Path: skills/content-plane/bolta.get_business_context/SKILL.md Purpose: Fetch workspace business DNA and context Inputs: workspace_id Outputs: Business DNA (industry, audience, positioning) Permissions: workspace:read Safe Mode: N/A (read-only)

bolta.get_voice_profile

Path: skills/content-plane/bolta.get_voice_profile/SKILL.md Purpose: Read voice profile configuration Inputs: voice_profile_id (optional, returns active if omitted) Outputs: Voice profile (tone, style, constraints) Permissions: voice:read Safe Mode: N/A (read-only)

bolta.list_recent_posts

Path: skills/content-plane/bolta.list_recent_posts/SKILL.md Purpose: List recent posts for context and continuity Inputs: account_id, limit, status filter Outputs: Array of recent posts with metadata Permissions: posts:read Safe Mode: N/A (read-only)

bolta.loop.from_template (deprecated)

Path: skills/content-plane/bolta.loop.from_template/SKILL.md Purpose: Generate multiple posts from a template Inputs: Template ID, count (1-50), variation parameters Outputs: Array of Post IDs Permissions: posts:write, templates:read Note: Deprecated in V2. Use bolta.agent.hire with template-based presets instead.

bolta.week.plan

Path: skills/content-plane/bolta.week.plan/SKILL.md Purpose: Plan a week's content with scheduling Inputs: Start date, posting frequency, themes Outputs: 7-day content calendar with scheduled posts Permissions: posts:write, posts:schedule Safe Mode: Routes to Pending Approval if ON Quota Impact: +5-15 to daily post count (spread across week)

Review Plane Skills

bolta.add_comment

Path: skills/review-plane/bolta.add_comment/SKILL.md Purpose: Add feedback comments to posts under review Inputs: post_id, comment text Outputs: Comment record Permissions: posts:read, posts:review Safe Mode: N/A (feedback operation)

bolta.approve_post

Path: skills/review-plane/bolta.approve_post/SKILL.md Purpose: Approve a draft post, moving it from Inbox to Approved status Inputs: draft_id, optional notes, optional schedule_time Outputs: Updated post status Permissions: posts:write, posts:approve Roles: Editor, Admin Safe Defaults: require_editor_role

bolta.reject_post

Path: skills/review-plane/bolta.reject_post/SKILL.md Purpose: Reject a draft with feedback for revision Inputs: draft_id, rejection_reason Outputs: Updated post status (Rejected) Permissions: posts:write, posts:review Roles: Editor, Admin

bolta.inbox.triage

Path: skills/review-plane/bolta.inbox.triage/SKILL.md Purpose: Organize pending posts by priority Inputs: Optional filters (platform, date range) Outputs: Categorized list of posts awaiting review Permissions: posts:read, posts:review Safe Mode: N/A (read-only)

bolta.review.digest

Path: skills/review-plane/bolta.review.digest/SKILL.md Purpose: Daily summary of posts awaiting review Inputs: None (workspace context) Outputs: Formatted summary with quick approve links Permissions: posts:read, posts:review Safe Mode: N/A (read-only) Note: Designed for cron execution (daily 9am)

bolta.review.approve_and_route

Path: skills/review-plane/bolta.review.approve_and_route/SKILL.md Purpose: Bulk approve and schedule posts Inputs: Post IDs or filter criteria Outputs: Updated post statuses Permissions: posts:write, posts:approve, posts:schedule Safe Mode: N/A (human override) Note: Bypasses Safe Mode (human decision)

Automation Plane Skills

bolta.cron.generate_to_review

Path: skills/automation-plane/bolta.cron.generate_to_review/SKILL.md Purpose: Daily content generation -> Pending Approval Inputs: N

…(truncated)