This skill should be used when performing content strategy analysis, keyword research, content auditing, content gap analysis, or AI agent consumability auditing. Sub-commands: auditing, planning, applying fixes.
16 stars
0 votes
0 copies
3 views
Added May 27, 2026
data-aishellbashgit
Works with
claude code
cli
Security analysis
B75/100
criticalSends environment variables or credentials to an external URL
Installs into .claude/skills of the current project.
Are you the author of Growth?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/jikig-ai-growth)
---
name: growth
description: "This skill should be used when performing content strategy analysis, keyword research, content auditing, content gap analysis, or AI agent consumability auditing. Sub-commands: auditing, planning, applying fixes."
---
<!-- soleur-cloud-mode:start -->
**Cloud Mode (Devin):** before pipeline work run `bash "${CLAUDE_PLUGIN_ROOT}/scripts/cloud-detect.sh"`. If `CLAUDE_PLUGIN_ROOT` is unset (cloud exec shells do not export it), resolve the root by IDENTITY, never by script basename: for `d` in `"$HOME/.local/share/devin/cli/plugins/cache"` and `/opt/.devin/plugins`, skip unless `[ -d "$d" ]`, then `MANIFEST="$(find "$d" -path '*/.claude-plugin/plugin.json' -exec grep -l '"name"[[:space:]]*:[[:space:]]*"soleur"' {} + 2>/dev/null | head -1)"`, `ROOT="${MANIFEST%/.claude-plugin/plugin.json}"` — one resolution, two consumers: `$ROOT/scripts/cloud-detect.sh` and `$ROOT/scripts/precommit-guard.sh`. (Shape check, not authentication: a planted `{"name":"soleur"}` dir passes — ADR-179 A11.) `local` or `not-local:no-devin-env` proceeds normally; any other `not-local:<reason>` applies `<plugin-root>/devin/INSTRUCTIONS.md` §Cloud Mode: emit the `--banner`, fan out sequentially with `Reviewed-Coverage: sequential-fallback` disclosure (never claim an independent review ran), `message_user` ack before any secrets read or production mutation, and run `precommit-guard.sh` before any `git commit` — hooks do not fire in cloud.
<!-- soleur-cloud-mode:end -->
<!-- grok-harness-invoke:start -->
**Grok Build (`plugins/soleur/lib/harness.ts` `invokeSkill()`):** Read this SKILL.md in this process and run it to completion. A one-segment `soleur:<name>` in this document names a SKILL — on Grok Build, Read `plugins/soleur/skills/<name>/SKILL.md` in this process; it is not a nested tool_use. A multi-segment id such as `soleur:<domain>:<name>` names an AGENT: spawn it, never Read it, and on Grok Build spawn_subagent takes the id with its colons replaced by hyphens (`agentIdToGrokSubagentType`). **Claude Code:** Skill tool for a skill (`soleur:<name>`), Task tool with `subagent_type` for an agent. Forbidden is executing a subset, not the Read.
<!-- grok-harness-invoke:end -->
# Growth Strategy
Analyze content for keyword alignment, plan content to capture search traffic, audit AI agent consumability, and apply content fixes. This skill delegates to the soleur:marketing:growth-strategist agent for content-level analysis and execution.
## Sub-commands
| Command | Description |
|---------|-------------|
| `growth audit <url-or-path>` | Audit existing content for keyword alignment and search intent match |
| `growth fix <path>` | Audit and apply keyword/copy/AEO fixes to local source files |
| `growth plan <topic> [--site <url-or-path>] [--competitors url1,url2]` | Research keywords, analyze gaps, and produce a prioritized content plan |
| `growth aeo <url-or-path>` | Audit content for GEO/AEO optimization (source citations, statistics, conversational readiness, FAQ structure, citation quality) |
If no sub-command is provided, display the table above and ask which sub-command to run.
---
## Sub-command: audit
Analyze existing site content for keyword alignment, search intent match, and readability. Produces a report with issues and rewrite suggestions.
### Steps
1. Parse the argument as a URL or local file/directory path. URLs start with `http://` or `https://`. Everything else is a local path.
2. Check for brand guide:
```bash
if [[ -f "knowledge-base/marketing/brand-guide.md" ]]; then
echo "Brand guide found. Will use for voice alignment."
fi
```
3. Launch the soleur:marketing:growth-strategist agent via the Task tool:
```
Task soleur:marketing:growth-strategist: "Audit the content at <url-or-path> for keyword alignment,
search intent match, and readability. <if brand guide exists: Also read
knowledge-base/marketing/brand-guide.md and align rewrite suggestions with the brand voice.>
Produce a structured report with per-page analysis, issues found, and rewrite suggestions.
Use WebFetch for URLs or Read/Glob for local paths."
```
4. Present the agent's report to the user.
---
## Sub-command: fix
Audit existing content and apply fixes to local source files. Combines analysis and execution in one pass, following the `seo-aeo fix` pattern.
### Steps
1. Parse the argument as a local file or directory path. If the argument starts with `http://` or `https://`, display: "growth fix works on local files only. For URL analysis, use `growth audit`." Stop execution.
2. Check for brand guide:
```bash
if [[ -f "knowledge-base/marketing/brand-guide.md" ]]; then
echo "Brand guide found. Will validate rewrites against brand voice."
fi
```
3. Launch the soleur:marketing:growth-strategist agent via the Task tool:
```
Task soleur:marketing:growth-strategist: "Audit the content at <path> for keyword alignment,
search intent match, readability, and GEO/AEO gaps. For each issue found, apply
a fix to the source files. Read each file before editing.
<if brand guide exists: Read knowledge-base/marketing/brand-guide.md and
validate all rewrites against the brand voice before applying.>
After all fixes, build the site to verify changes compile.
Report what was changed per file."
```
4. Present the agent's report showing changes made.
5. If the build failed, show the error and suggest: "Run `git checkout -- <file>` to revert changes, or fix the build error manually."
---
## Sub-command: plan
Research keywords, analyze content gaps, and produce a prioritized content plan. This is a self-contained workflow that combines keyword research, gap analysis, and planning.
### Steps
1. Parse arguments:
- `<topic>` (required): the topic or keyword space to research
- `--site <url-or-path>` (optional): the user's existing site for gap analysis
- `--competitors url1,url2` (optional): competitor URLs for comparison
2. Check for brand guide (same as audit step 2).
3. Launch the soleur:marketing:growth-strategist agent via the Task tool:
```
Task soleur:marketing:growth-strategist: "Create a content plan for the topic '<topic>'.
1. Research keywords related to this topic using WebSearch. Classify each by search
intent (informational, navigational, commercial, transactional) and relevance.
2. <if --site provided: Analyze the existing content at <site> to identify gaps --
topics where the site has no coverage or only partial coverage.>
3. <if --competitors provided: Fetch competitor sites at <urls> via WebFetch and
compare their content coverage against the target keywords. Skip unreachable
competitors and note them.>
4. Produce a prioritized content plan: P1 (high impact), P2 (medium), P3 (future).
Each piece should include content type, target keywords, search intent, and outline.
<if brand guide exists: Read knowledge-base/marketing/brand-guide.md and align
keyword relevance and content priorities with the brand positioning.>"
```
4. Present the agent's report to the user.
---
## Sub-command: aeo
Audit content for AI agent consumability at the content level. Checks whether AI models can accurately extract, cite, and quote the content.
### Steps
1. Parse the argument as a URL or local path (same heuristic as audit).
2. Launch the soleur:marketing:growth-strategist agent via the Task tool:
```
Task soleur:marketing:growth-strategist: "Audit the content at <url-or-path> for GEO/AEO optimization.
Check source citation presence, statistics/specificity, conversational readiness,
FAQ structure quality, definition extractability, summary quality, and
citation-friendly paragraph structure.
Do NOT check JSON-LD, meta tags, sitemaps, or llms.txt format -- those belong to
the soleur:marketing:seo-aeo-analyst agent. Focus only on content-level checks.
Use WebFetch for URLs or Read/Glob for local paths."
```
3. Present the agent's report to the user.
4. If technical AEO checks are also needed, suggest: "Run `seo-aeo audit` for technical SEO and AEO checks (JSON-LD, meta tags, llms.txt)."
---
## Important Guidelines
- Each sub-command is independent. No sub-command requires a prior run of another.
- The `plan` sub-command performs its own keyword research internally -- no need to run a separate research step first.
- The `audit`, `plan`, and `aeo` sub-commands produce inline output only. The `fix` sub-command modifies source files directly.
- The soleur:marketing:growth-strategist agent handles content-level analysis. For technical SEO (meta tags, JSON-LD, sitemaps, llms.txt), direct users to the `seo-aeo` skill instead.
- For generating new articles from content plans, direct users to the `content-writer` skill.