Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/stepolan/marketing-repo-templateWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/stepolan/marketing-repo-template/extract-voice)<a href="https://agentmods.dev/commands/stepolan/marketing-repo-template/extract-voice"><img src="https://agentmods.dev/badge/commands/stepolan/marketing-repo-template/extract-voice/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/stepolan/marketing-repo-template/extract-voice"><img src="https://agentmods.dev/badge/commands/stepolan/marketing-repo-template/extract-voice.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00010 | $0.00691 |
| Opus 5 | $0.00005 | $0.00345 |
| Sonnet 5 | $0.00002 | $0.00138 |
| Haiku 4.5 | $0.00001 | $0.00069 |
Grade A, and why
extract-voice scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Analyze an author's existing content and generate a voice profile from it.
Arguments: $ARGUMENTS (the author name, matching a directory under authors/)
Execute the following steps:
-
Find all content by this author:
- Scan
authors/<name>/posted/linkedin/for published posts - Scan
authors/<name>/drafts/linkedin/for drafts - Scan
corporate/drafts/blog/for blog posts where the Author field matches - Scan
authors/<name>/samples/for any imported prior work - If no content found, tell the user to add existing content first
- Scan
-
Read every file found. You need volume to extract patterns. Read all of them.
-
Analyze the writing for these dimensions:
Sentence and paragraph structure:
- Average sentence length
- How paragraphs are constructed (short/punchy vs. flowing)
- Use of fragments vs. complete sentences
Rhetorical patterns:
- How do they open a piece? (Question, bold claim, scenario, statistic)
- How do they close? (Question, call to action, reflection)
- Do they use rhetorical questions? How often?
- Do they use lists, frameworks, or narrative arcs?
Tone and stance:
- Formal vs. conversational
- Authoritative vs. exploratory
- Contrarian vs. consensus-building
- How do they handle disagreement or critique?
Word-level patterns:
- Contractions or no contractions
- Punctuation habits (em dashes, semicolons, ellipses, exclamation marks)
- Jargon level (technical terms, industry shorthand, plain language)
- Filler or hedge words they use or avoid
- Characteristic phrases or constructions that recur
Content patterns:
- Do they use data and citations?
- Do they tell stories or paint scenarios?
- Do they reference personal experience?
- Do they name-drop or give credit to others?
-
Generate a voice profile following the structure in the existing template at
authors/<name>/voice/voice-profile.md. The output must include:- Core Characteristics (bulleted list of defining traits)
- Voice Patterns (with named patterns and examples pulled from their actual writing)
- Words/Phrases to AVOID (patterns you did NOT see and that would sound off-voice)
- Formatting Rules (what you observed about structure)
- Instructions for AI (numbered rules for reproducing this voice)
- Voice Check (checklist of does/does-not sound like this author)
-
Show the profile to the user for review before writing it. Ask if anything is wrong or missing.
-
After approval, write it to
authors/<name>/voice/voice-profile.md. -
Report:
- How many pieces were analyzed
- The strongest patterns (most consistent across pieces)
- Any inconsistencies noticed (voice shifts between pieces that may need a decision)
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 71 lines · 10 tokens per session scan A b1f751452a15
extract-voice is a command published in the GitHub repository stepolan/marketing-repo-template (2 stars, last pushed 4mo ago), licensed MIT. It adds 10 tokens to every session and 691 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other commands, from other repositories
blog-assets
Build HTML templates for each visual placeholder in a blog article and batch-export them as 1200x675 PNGs at 2x. Fourth step in the blog production pipeline.
repurpose
Fan a QA'd long-form piece (blog, research, weekly report) down into 6 channel-native formats — X thread, X singles, LinkedIn article, newsletter snippet, and pull-quote cards. Every output is canon-checked and staged for review.
blog-pipeline
Reference document showing the full Blog Engine flow with human approval gates. This is not an executable command — it's a map of how blog commands connect.
compliance-check
Jurisdiction-aware content compliance validation. Scans copy for regulatory violations, banned claims, missing disclaimers, and financial promotion language.
blog-publish
Push a completed blog article to the CMS -- upload images, create CMS item, sync components. Final step in the blog production pipeline.
community-harvest
Mine community channels (Discord, Slack, Telegram) for recurring questions, pain points, and feature requests. Produces actionable content briefs and documentation gap reports.