ai-content-pipelinelisted

# AI Content Pipeline Use this skill when the job spans multiple stages of the TN content system. ## Primary Owners - `sadhana-orchestrator` - `pi` as upstream brief support ## Canonical OpenClaw Pipeline 1. Source material or published post 2. Content Engine brief 3. Distribution packet 4. Creative staging or generative asset production 5. Dry-run or live execution 6. Receipt capture and dashboard visibility ## Bound Adapters - Blog or source -> packet: ```bash python3 <openclaw-scripts>/blog_to_distribution_packet.py --slug <post-slug> ``` - Stage article-native assets: ```bash python3 <openclaw-scripts>/stage_distribution_assets.py --packet /abs/path/to/packet.json --overwrite ``` - Generate Higgsfield assets when required: ```bash python3 <openclaw-scripts>/higgsfield_generate.py --packet /abs/path/to/packet.json ``` - Execute inference dry-run or publish: ```bash python3 <openclaw-scripts>/inference_flow_run.py --packet /abs/path/to/packet.json --dry-run ``` - Hand off actionable work to Paperclip: ```bash python3 <openclaw-scripts>/handoff-to-paperclip.py --packet /abs/path/to/packet.json ``` - Review live state: `http://127.0.0.1:5199/api/distribution/overview` ## Guardrails - X is the source lane. - Reddit and Substack are manual expansions. - Threads, Instagram, and Facebook are repost layers. - Do not bypass `manual-confirmed` publish policy. - Use the adapters above instead of raw provider calls whenever possible.