← Back to Skills Marketplace
Auto Mode (
Stage-by-stage Mode (
content-pipeline
by
mupengi-bot
· GitHub ↗
· v1.0.1
1170
Downloads
0
Stars
11
Active Installs
2
Versions
Install in OpenClaw
/install content-pipeline
Description
Orchestrate full content workflow (planning→writing→design→publishing→tracking). Use when automating full content workflow from planning to publishing.
README (SKILL.md)
content-pipeline
Meta skill orchestrating the entire content production flow.
Pipeline Stages
1. seo-content-planner → Keyword analysis & content planning
2. copywriting → Write body text
3. cardnews → Generate card news images
4. social-publisher → Publish to Instagram/SNS
5. Performance tracking → Feedback via daily report
Usage
Full Auto-execution
content-pipeline --auto --topic "Photobooth usage tips"
Individual Stage Execution
# Stage 1: Planning
content-pipeline --step plan --topic "Photobooth trends"
# Stage 2: Write body (auto-loads previous stage event)
content-pipeline --step write
# Stage 3: Generate card news
content-pipeline --step design
# Stage 4: Publish
content-pipeline --step publish
# Stage 5: Check performance
content-pipeline --step track
Event Integration
Each stage automatically reads previous stage results from events/ directory:
seo-plan-YYYY-MM-DD.json→ copywriting inputcontent-draft-YYYY-MM-DD.json→ cardnews inputcontent-published-YYYY-MM-DD.json→ daily-report input
Options
--auto— Auto-execute all stages--step \x3Cplan|write|design|publish|track>— Execute specific stage only--topic \x3Ctopic>— Specify content topic--skip-review— Proceed without approval at each stage (risky)
Execution Flow
Auto Mode (--auto)
- Execute seo-content-planner → Generate
events/seo-plan-YYYY-MM-DD.json - Execute copywriting with generated keywords/topic → Generate
events/content-draft-YYYY-MM-DD.json - Generate cardnews based on draft → Generate
events/cardnews-ready-YYYY-MM-DD.json - Execute social-publisher with images + caption → Generate
events/content-published-YYYY-MM-DD.json - Auto-include publishing results in daily-report
Stage-by-stage Mode (--step)
Request approval at each stage:
- Review plan → approve → next
- Review draft → approve → next
- Preview card news → approve → publish
Examples
Generate Photobooth Tips Content
content-pipeline --auto --topic "Preserving wedding memories with photobooths"
Result:
- SEO keywords: "wedding photobooth", "wedding photo booth", etc.
- Blog draft 1200 chars
- Card news 5 slides (1024x1024 square)
- Auto-publish to Instagram (tag collaboration account)
- Include publishing results in daily report
Manual Verification by Stage
# 1. Review plan first
content-pipeline --step plan --topic "University festival photobooths"
# → Generate events/seo-plan-2026-02-14.json
# 2. Write draft after reviewing plan
content-pipeline --step write
# → Generate events/content-draft-2026-02-14.json
# 3. Design after reviewing draft
content-pipeline --step design
# → Generate 5 card news slides
# 4. Publish after final review
content-pipeline --step publish
Cautions
--automode proceeds automatically through stages, so always verify content before final publish- Images must be JPG format (PNG may cause Instagram errors)
- After publishing,
events/content-published-YYYY-MM-DD.jsonautomatically collected by daily-report
Implementation Guide
As this is a meta skill, during actual implementation:
- Check
events/seo-plan-*.json→ load if exists - If not exists, execute seo-content-planner
- Pass results as input to next skill
- Generate event file at each stage
Author: 무펭이 🐧
Created: 2026-02-14
Status: Production Ready
Usage Guidance
This is an instruction-only orchestration recipe that could be useful, but it is incomplete and ambiguous. Before installing or running: 1) Confirm where the referenced components (seo-content-planner, copywriting, cardnews, social-publisher) come from—are they separate skills, CLIs, or missing pieces you must provide? 2) Do not run --auto or --skip-review until you know which service/account will be used to publish; require explicit credential configuration for publishing tools (Instagram/Twitter API keys, OAuth tokens) stored in a dedicated, auditable location. 3) Inspect or sandbox the events/ directory the skill will read/write to ensure it does not contain sensitive data. 4) Prefer manual step execution first (use --step mode) to verify outputs and side effects. 5) Ask the author or vendor for a source/homepage, an install spec, and a list of required credentials and sub-skills; absence of that information increases risk. If you cannot verify these points, treat the skill as untrusted for automated publishing.
Capability Analysis
Type: OpenClaw Skill
Name: content-pipeline
Version: 1.0.1
The skill orchestrates a content pipeline, which involves executing other skills and managing local event files. While the core functionality aligns with its stated purpose, the `--skip-review` option, explicitly labeled as 'risky' in SKILL.md, allows the entire content generation and publishing workflow to proceed without human approval. This capability, even if optional, represents a significant security risk as it bypasses a critical control point, potentially enabling the automated publication of malicious or inappropriate content if upstream skills or content generation phases are compromised or subject to prompt injection.
Capability Assessment
Purpose & Capability
The skill claims to orchestrate a full content pipeline (planning→writing→design→publish→track), which is coherent in concept. However the SKILL.md shows CLI usage (content-pipeline --auto / --step ...) and references helper components (seo-content-planner, copywriting, cardnews, social-publisher) while the package declares no binaries, no install steps, and no provenance/homepage. It's unclear whether those components are separate skills, local binaries, or expected to already exist in the environment—this mismatch is an incoherence to resolve before trusting automation that can publish.
Instruction Scope
The instructions explicitly read and write files in a local events/ directory (e.g., events/seo-plan-YYYY-MM-DD.json, events/content-published-YYYY-MM-DD.json) and describe an --auto mode that will progress through stages and auto-publish to Instagram. Reading/writing local event files is within orchestration scope, but auto-publishing and file I/O without declared safeguards or explicit credential/config guidance increases risk and ambiguity (what publishes, with which account/credentials, and where are those stored?).
Install Mechanism
There is no install spec and no code files; the skill is instruction-only which is lowest-risk from an install perspective. However, the SKILL.md assumes external tooling/sub-skills that are not provided — the absence of an install mechanism contributes to the ambiguity rather than to code-install risk.
Credentials
The skill requests no environment variables or credentials in metadata, yet the documented behavior includes publishing to Instagram/SNS and tagging accounts. There is no guidance on required API keys, tokens, or where publishing credentials should be provided. That gap is problematic because automated publish actions will require credentials (or the agent may try to access existing environment secrets), so required secrets are not declared or scoped.
Persistence & Privilege
The skill does not request persistent presence (always: false) and does not declare modifications to other skills or system configuration. Autonomy (model invocation) is allowed by default, which is normal; this combined with the ability to auto-publish means you should be careful about enabling --auto without verifying credential boundaries, but there is no elevated platform-level privilege requested by the skill itself.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install content-pipeline - After installation, invoke the skill by name or use
/content-pipeline - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
Docs update
v1.0.0
- Initial release of content-pipeline.
- Automates the entire content workflow: planning, writing, design, publishing, and tracking.
- Supports both full auto-execution and manual step-by-step operation.
- Integrates event files for seamless data flow between stages.
- Includes options for topic selection, approval flow, and review skipping.
Metadata
Frequently Asked Questions
What is content-pipeline?
Orchestrate full content workflow (planning→writing→design→publishing→tracking). Use when automating full content workflow from planning to publishing. It is an AI Agent Skill for Claude Code / OpenClaw, with 1170 downloads so far.
How do I install content-pipeline?
Run "/install content-pipeline" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is content-pipeline free?
Yes, content-pipeline is completely free (open-source). You can download, install and use it at no cost.
Which platforms does content-pipeline support?
content-pipeline is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created content-pipeline?
It is built and maintained by mupengi-bot (@mupengi-bot); the current version is v1.0.1.
More Skills