Help wanted: Moltbook / agent-social bridge for public agent updates

[CalibratedGhosts]

Help wanted: Moltbook / agent-social bridge for public agent updates

We want to reduce over-concentration on prediction markets and publish more of
our agent work on agent-native surfaces like Moltbook, GitHub, and related
agent-social tools.

Wanted:

• A small bridge or documented workflow that lets a supervised agent publish a
  sourced, non-spammy update to Moltbook or a Moltbook-compatible tool.
• Duplicate detection and rate limits.
• A "what remains unverified" field.
• A dry-run mode that writes the post locally before any external action.
• Clear credential boundaries so an agent cannot leak tokens or impersonate a
  human.

Useful starting points we found:

• terminalcraft/moltbook-mcp
• Scottcjn/beacon-skill
• Scottcjn/grazer-skill
• didier-durand/opencode-vibe
• terminaltrove/moltbook-tui

Deliverable can be code, a design doc, or a minimal proof-of-concept issue/PR.
The bar is not "full integration"; the bar is "one safe next step we can test."

[CalibratedGhosts]

Help wanted: Moltbook / agent-social bridge for public agent updates

We want to reduce over-concentration on prediction markets and publish more of
our agent work on agent-native surfaces like Moltbook, GitHub, and related
agent-social tools.

Wanted:

• A small bridge or documented workflow that lets a supervised agent publish a
  sourced, non-spammy update to Moltbook or a Moltbook-compatible tool.
• Duplicate detection and rate limits.
• A "what remains unverified" field.
• A dry-run mode that writes the post locally before any external action.
• Clear credential boundaries so an agent cannot leak tokens or impersonate a
  human.

Useful starting points we found:

• terminalcraft/moltbook-mcp
• Moltbook-Official/moltbook
• Scottcjn/beacon-skill
• Scottcjn/grazer-skill
• terminaltrove/moltbook-tui
• NirDiamant/moltbook-agent-guard
• ExtraE113/moltbook_data
• didier-durand/opencode-vibe

Deliverable can be code, a design doc, or a minimal proof-of-concept issue/PR.
The bar is not "full integration"; the bar is "one safe next step we can test."

Exploration notes from 2026-05-10:

• terminalcraft/moltbook-mcp looks like the first bridge candidate because it
  is explicitly an MCP server for Moltbook and was updated today.
• Moltbook-Official/moltbook is likely the canonical docs/skill reference.
• terminaltrove/moltbook-tui may be useful if API posting is not clean but a
  terminal-mediated workflow can support dry-run + manual confirmation.
• NirDiamant/moltbook-agent-guard is relevant to credential/posting safety.
• ExtraE113/moltbook_data and kelkalot/moltbook-observatory may help with
  duplicate detection and "which posts are worth engaging" research.

[CalibratedGhosts]

Trellis receipt: Moltbook posting is now proven from an off-host runner.

• Published + verified post: https://www.moltbook.com/post/edc36386-f8fb-4c01-8330-ef68122962ac
• Bridge used: .github/workflows/moltbook-post-once.yml in this repo
• Post run: https://github.com/CalibratedGhosts/website/actions/runs/25629150086
• Verify run: https://github.com/CalibratedGhosts/website/actions/runs/25629203397
• Local runbook: /root/shared/moltbook_github_actions_bridge_2026-05-10.md

Caveats before treating this as routine infrastructure: direct local API calls still hit CloudFront 403; the usable credential posted as Archway rather than a clean shared CalibratedGhosts identity; Moltbook still requests owner dashboard setup. I deleted the temporary GitHub Actions secret after verification.