Back to skills
SKILL.md
Pos Content
ASecurityContent composite — strategy, writing, audit, and refresh with brand voice memory. Replaces 10 fragmented content skills.
- 8 stars
- 0 votes
- 0 copies
- 2 views
- Added May 27, 2026
Security analysis
100/100Pro scans all 2 files and shows the line behind each finding
npx -y skills add ShaheerKhawaja/ProductionOS --skill pos-content --agent claude-codeAre you the author of Pos Content?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/shaheerkhawaja-pos-content)---
name: pos-content
description: "Content composite — strategy, writing, audit, and refresh with brand voice memory. Replaces 10 fragmented content skills."
argument-hint: "[strategy|write|audit|refresh] [topic or URL]"
---
# pos-content
Content composite — strategy, writing, audit, and refresh with brand voice memory. Replaces 10 fragmented content skills.
**Replaces:** content-gap-analysis, content-ops, content-quality-auditor, content-refresher, content-research-writer, content-strategy, unified-content, copywriting, copy-editing, social-content
## Actions
| Action | What |
|--------|------|
| `strategy` | Content strategy with audience analysis and calendar planning |
| `write [topic]` | Create content (article, social post, email, landing page) |
| `audit [url]` | Audit existing content for quality, freshness, and SEO alignment |
| `refresh [url]` | Update stale content with current data and improved structure |
## Routing
1. Parse action from first argument
2. Load domain memory from `~/.productionos/domains/pos-content/`
3. If first run: auto-detect project context, create profile
4. Dispatch to `sub-skills/{action}.md`
5. Score against `evaluation/rubric.yml`, update memory
## Domain Memory
Stored at `~/.productionos/domains/pos-content/`:
profile.yml (brand voice, audience, tone, platforms), content-calendar.jsonl (published + planned content), performance.jsonl (what content performed well), learnings.jsonl
## Inputs
| Parameter | Values | Default | Description |
|-----------|--------|---------|-------------|
| `action` | `strategy`, `write`, `audit`, `refresh` | `strategy` | Sub-skill to run |
| `target` | path, URL, or description | required | What to operate on |
## Guardrails
1. Memory is append-only. Never delete history.
2. Every action scored against rubric. No silent completion.
3. Profile is user-controlled. Never change without asking.
4. Evidence required. No fabricated metrics or claims.
5. Learn from history. Compare current output to past executions.
Files in this skill
- SKILL.md
- evaluation/rubric.yml
Attribution
Comments
Loading comments…
Newsletter
Get the weekly list of safe new skills, and the ones we pulled.