What does this skill do, and when should you use it?
The postiz skill exposes the Postiz CLI to AI agents for scheduling and managing posts across 28+ social platforms. It covers the full workflow: authentication, integration discovery, media upload, publishing, draft management, platform-specific settings, and analytics. All commands emit JSON, making them easy for agents to parse and chain. The repo also ships as a Claude Code plugin, a Cursor plugin, and a Grok plugin marketplace entry, installed via npx skills.
Runs the postiz CLI: auth:login/auth:status (OAuth2 device flow or POSTIZ_API_KEY) for authentication; integrations:list and integrations:settings to discover connected accounts and their settings schemas; postiz upload to push media and get Postiz-verified URLs (required by TikTok, Instagram, YouTube and others); posts:create to publish scheduled multi-platform posts with media, comments/threads, and per-provider settings; posts:list/status/delete/settings for post management; analytics:platform and analytics:post for metrics, with posts:missing + posts:connect to resolve {"missing": true} responses when platforms don't return a post ID.
- A social media manager delegating a weekly multi-platform campaign with identical content to X, LinkedIn, and Facebook.
- A video creator publishing to TikTok with DIRECT_POST after uploading the video via postiz upload.
- A community admin posting to a specific subreddit with the correct flair fetched dynamically.
- A marketing team batch-scheduling a week of content and truncating copy to each platform's character limit.
- An agent user monitoring followers, impressions, and engagement, and fixing missing release IDs when analytics return {"missing": true}.
How do you install this skill?
- Publishing/deleting posts are irreversible external effects; verify with draft mode first and add your own confirmation step
- Skill grants broad Bash(postiz:*) permission — review the authorization scope carefully
- Core function depends on postiz.com, npm, and overseas social platforms; mainland-China reachability unverified; no Chinese support
- License metadata is NOASSERTION and publisher unverified — confirm licensing yourself
- Docs show version inconsistency and a truncated SKILL.md tail; delay units (seconds vs minutes) are inconsistent across examples
- Static review executed nothing — actual usability and silent-failure risks (TikTok etc.) need your own testing
- Shell / CLI
- Network access
- Local filesystem
Node.js (npm/pnpm)postiz CLI (npm install -g postiz)
1) Install the CLI: npm install -g postiz (or pnpm install -g postiz). 2) Install the skill: npx skills add gitroomhq/postiz-agent. Or as a Claude Code plugin: /plugin marketplace add gitroomhq/postiz-agent then /plugin install postiz@postiz-agent. Cursor users can install from the Customize panel; Grok Build users can install from the xAI plugin marketplace (the Grok plugin also bundles the hosted MCP server at mcp.postiz.com). Manual local installs are described via the repo's .cursor-plugin/plugin. manifest.
git clone --depth 1 https://github.com/gitroomhq/postiz-agent.git ~/.claude/skills/postiz-agentGenerated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Authenticate first: postiz auth:login (device flow) or export POSTIZ_API_KEY=your_api_key; optionally export POSTIZ_API_URL for a self-hosted API. Then follow the workflow: postiz auth:status → postiz integrations:list for integration IDs → postiz integrations:settings <id> for rules and schemas → postiz upload <file> and extract .path with jq → postiz posts:create -c "Content" -s "2024-12-31T12:00:00Z" -i "integration-id" -m "uploaded-url". Caveats: values passed to -m must be the .path from postiz upload, never raw files or external URLs; TikTok requires content_posting_method set to DIRECT_POST; the schedule date is a required ISO 8601 timestamp.
What are this skill's strengths and limitations?
- Covers 28+ platforms including X, LinkedIn, Reddit, YouTube, TikTok, Instagram, and Facebook
- All commands output JSON, easy for agents to parse programmatically
- OAuth2 device-flow authentication with no manual client-secret handling
- Integration tools workflow dynamically fetches Reddit flairs, YouTube playlists, and similar data
- SKILL.md documents common pitfalls explicitly (mandatory media upload, TikTok silently discarding invalid settings)
- Hard dependency on the Postiz service and Postiz API (default api.postiz.com) — not a purely local tool
- The GitHub repo's License field is NOASSERTION while the main Postiz project is AGPL-3.0; verify before commercial use
- Media not passed through postiz upload is rejected by most platforms, adding a mandatory step
- Some platforms (e.g. TikTok) don't return a post ID after publishing, requiring manual posts:connect before analytics work
- No independent test suite or per-platform verification beyond the repo's own examples
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Postiz Agent Skill this page | 52 · Use with care | ★ 517 | 11d ago | NOASSERTION |
| Upload-Post Social Media Publishing Skill | 47 · Use with care | ★ 38 | 3mo ago | — |
| Authsome | 46 · Use with care | ★ 98 | 3mo ago | MIT |
| DOCA Telemetry Utils ✓ NVIDIA · Official | 56 · Use with care | ★ 3.6k | 11d ago | Apache-2.0 |
| TAO on Brev GPU Instances ✓ NVIDIA · Official | 51 · Use with care | ★ 3.6k | 11d ago | Apache-2.0 |
How did FollowSkills review this skill?
Docs transparently disclose auth (OAuth2/key), credential storage (~/.postiz/credentials.), API key via env vars, and data flow (upload to Postiz CDN then publish); no concealment or malicious behavior. Deducted: broad Bash(postiz:*) grant; publishing/deleting posts are irreversible external effects with no built-in user-confirmation gate in SKILL.md; draft mode mentioned but not enforced; publisher unverified.
SKILL.md is highly self-consistent with four hard rules, gotchas, error-handling and retry scripts, and clear failure feedback (e.g., TikTok silent setting discard). Deducted: static review cannot reproduce key paths; no committed test evidence covering the skill; SKILL.md is truncated mid-list at the end; delay-unit inconsistency (-d 2000 seconds vs -d 5 minutes) invites misreading.
Audience (developers/marketers scheduling posts via AI agents) and trigger scenarios are clear, 28+ platforms enumerated, parameters and outputs exemplified. Deducted: core function depends entirely on postiz.com, npm, and overseas social platforms with unknown mainland-China reachability; no Chinese-language support; limited disclosure of recovery boundaries after upload failures.
Well-layered docs (SKILL.md→deep dives→examples), Keep a Changelog format with version numbers, complete install notes. Deducted: license metadata NOASSERTION; version inconsistency across docs (PROJECT_STRUCTURE says 1.0.0, CHANGELOG at 2.0.17); truncated SKILL.md tail; publisher maintenance responsibility/update path not explicitly stated.
Core task (scheduling social posts via CLI) is fully described and output would be directly usable, with clear marginal value over manual posting; draft/retry modes control cost. Deducted: static review cannot verify real output usability; no executed representative outputs; value claim partly leans on promotional recommendations of third-party CLIs like agent-media.
Substantial auditable primary material (command syntax, API endpoints, JSON schemas, response examples) enabling cross-checking. Deducted: all author-authored; no independent reproduction, third-party execution evidence, or test results; marketing-style summaries (✅ fully supported) mixed with factual claims.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →