config.yaml¶

Infrastructure configuration for social-hook. Controls models, platforms, scheduling, channels, and media generation.

Location¶

Priority	Location
1	`--config` / `-c` CLI flag
2	`SOCIAL_HOOK_CONFIG` environment variable
3	`~/.social-hook/config.yaml` (default)

Interactive setup¶

social-hook setup                  # Full wizard
social-hook setup --only apikeys   # Just API keys
social-hook setup --only platforms # Just platforms
social-hook config show            # View current config
social-hook config show --content  # View content-config.yaml
social-hook config get models.evaluator
social-hook config set scheduling.timezone America/Los_Angeles
social-hook config set context.max_discovery_tokens 80000 --content  # Edit content config

models¶

LLM model selection. Format: provider/model-id.

Known providers: anthropic, claude-cli, openai, openrouter, ollama.

Field	Default	Description
`evaluator`	`anthropic/claude-opus-4-5`	Evaluates commits, makes draft/skip decisions, runs discovery
`drafter`	`anthropic/claude-opus-4-5`	Drafts social media content
`gatekeeper`	`anthropic/claude-haiku-4-5`	Lightweight checks, project summaries

models:
  evaluator: anthropic/claude-opus-4-5
  drafter: anthropic/claude-opus-4-5
  gatekeeper: anthropic/claude-haiku-4-5

platforms¶

Each key is a platform name. Built-in platforms: x, linkedin. Custom platform names are also supported.

Field	Type	Default	Description
`enabled`	bool	`false`	Whether this platform is active
`priority`	string	`"secondary"`	`"primary"` or `"secondary"` — affects smart defaults
`type`	string	auto	`"builtin"` for x/linkedin, `"custom"` otherwise
`account_tier`	string	—	X-specific: `"free"` (280 chars), `"basic"`, `"premium"`, `"premium_plus"` (25000 chars)
`description`	string	—	Extra context for the drafter (custom platforms)
`format`	string	—	Output format hint (e.g., `"tweet"`, `"post"`, `"article"`)
`max_length`	int	—	Character limit. `null` = no limit
`filter`	string	smart default	Content filter: `"all"`, `"notable"`, `"significant"`
`frequency`	string	smart default	Posting frequency: `"high"`, `"moderate"`, `"low"`, `"minimal"`
`scheduling`	object	—	Per-platform scheduling overrides

Smart defaults¶

When filter and frequency are omitted, they resolve based on platform name and priority:

Platform	Primary	Secondary
x	filter=`all`, freq=`high`	filter=`notable`, freq=`moderate`
linkedin	filter=`notable`, freq=`moderate`	filter=`significant`, freq=`low`
custom	filter=`notable`, freq=`moderate`	filter=`significant`, freq=`low`

Content filters¶

Filter	Episode types that pass
`all`	Everything
`notable`	milestone, launch, synthesis, demo_proof, before_after, postmortem
`significant`	milestone, launch, synthesis

Frequency presets¶

Frequency	max_posts_per_day	min_gap_minutes
`high`	3	30
`moderate`	1	120
`low`	1	2880 (2 days)
`minimal`	1	10080 (7 days)

platforms:
  x:
    enabled: true
    priority: primary
    account_tier: free
  linkedin:
    enabled: true
    priority: secondary

scheduling¶

Global scheduling defaults. Per-platform overrides take precedence.

Field	Type	Default	Description
`timezone`	string	`"UTC"`	IANA timezone (e.g., `"America/Los_Angeles"`)
`max_posts_per_day`	int	`3`	Maximum posts per day
`min_gap_minutes`	int	`30`	Minimum gap between posts
`optimal_days`	list	`["Tue", "Wed", "Thu"]`	Preferred posting days
`optimal_hours`	list	`[9, 12, 17]`	Preferred posting hours (24h)
`max_per_week`	int	`10`	Maximum posts per week
`thread_min_tweets`	int	`4`	Minimum tweets in an X thread

scheduling:
  timezone: America/New_York
  max_posts_per_day: 2
  min_gap_minutes: 60
  optimal_days: [Mon, Tue, Wed, Thu, Fri]
  optimal_hours: [9, 12, 17]

media_generation¶

Controls which media tools are available for draft content.

Field	Type	Default	Description
`enabled`	bool	`true`	Master toggle for all media generation
`tools.mermaid`	bool	`true`	Mermaid diagram generation
`tools.nano_banana_pro`	bool	`true`	AI image generation (Google Gemini)
`tools.playwright`	bool	`true`	Browser screenshot capture
`tools.ray_so`	bool	`true`	Code screenshot generation

media_generation:
  enabled: true
  tools:
    mermaid: true
    nano_banana_pro: true
    playwright: true
    ray_so: true

Note

Tool availability is set here. Content guidance for how to use each tool is in content-config.yaml.

channels¶

Messaging channels for notifications and bot interaction. Keys: telegram, slack, web.

Field	Type	Default	Description
`enabled`	bool	`false`	Enable this channel
`allowed_chat_ids`	list	`[]`	Authorized chat IDs (Telegram/Slack)

The web channel is enabled by default.

channels:
  web:
    enabled: true
  telegram:
    enabled: true
    allowed_chat_ids: ["123456789"]

consolidation¶

Batches rapid-fire commits into consolidated evaluations.

Field	Type	Default	Description
`enabled`	bool	`false`	Enable consolidation processing
`mode`	string	`"notify_only"`	`"notify_only"` or `"re_evaluate"`
`batch_size`	int	`20`	Max decisions per consolidation batch
`auto_consolidate_drafts`	bool	`true`	Auto-consolidate draft decisions
`consolidate_approved`	bool	`false`	Include approved drafts in consolidation
`time_window_hours`	float	`4.0`	Time window for grouping
`time_window_max_drafts`	int	`3`	Max drafts per time window

journey_capture¶

Development Journey capture via Claude Code hooks.

Field	Type	Default	Description
`enabled`	bool	`false`	Enable journey capture
`model`	string	—	Override model for capture. `null` = use `models.evaluator`

notification_level¶

Field	Type	Default	Description
`notification_level`	string	`"all_decisions"`	`"all_decisions"` or `"drafts_only"`

Environment variables (.env)¶

Stored in ~/.social-hook/.env. Real environment variables override file values.

Variable	Required for	Description
`ANTHROPIC_API_KEY`	`anthropic/` models	Claude API key
`OPENAI_API_KEY`	`openai/` models	OpenAI API key
`OPENROUTER_API_KEY`	`openrouter/` models	OpenRouter API key
`OLLAMA_BASE_URL`	`ollama/` models	Ollama server URL
`X_CLIENT_ID`	X platform	OAuth 2.0 client ID
`X_CLIENT_SECRET`	X platform	OAuth 2.0 client secret
`LINKEDIN_CLIENT_ID`	LinkedIn	OAuth client ID
`LINKEDIN_CLIENT_SECRET`	LinkedIn	OAuth client secret
`LINKEDIN_ACCESS_TOKEN`	LinkedIn	Access token
`TELEGRAM_BOT_TOKEN`	Telegram channel	Bot token
`TELEGRAM_ALLOWED_CHAT_IDS`	Telegram channel	Comma-separated chat IDs
`GEMINI_API_KEY`	nano_banana_pro media	Google Gemini API key