# ChunkMaps Skills Library

> Source: [https://chunkmaps.com/NLP_improved_communication/chunkmap-skills.html](https://chunkmaps.com/NLP_improved_communication/chunkmap-skills.html)

Browse the ChunkMaps skills library: persuasion flow, agent blackboard, and policy-to-pipeline playbooks for humans and developer agents-with HTML guides, raw .md files, and a copyable README.

Skills library · Hub

ChunkMaps skills are editable **.md playbooks** for humans and developer agents. Each file captures when to use a pattern, required inputs, step-by-step procedure, output schema, and guardrails-so method knowledge stays out of hard-coded app logic.

## Browse skills

  [How-toHow to use .md skill files](how_to_use_skill_files.html)
  [v1 · PersuasionPersuasion Flow guide](chunkmap_persuasion_flow.html)
  [Raw .mdPersuasion Flow skill](skills/chunkmap-persuasion-flow.md)
  [AgenticOrchestrators guide](agent_orchestrators_chunkmaps.html)
  [Raw .mdAgent Blackboard skill](skills/chunkmap-agent-blackboard.md)
  [AgenticEnterprise guide](enterprise_architects_chunkmaps.html)
  [Raw .mdPolicy to Pipeline skill](skills/chunkmap-policy-to-pipeline.md)
  [PromptsRaw paste-ready prompts](skills/raw-prompts/README.md)
  [SpecJSON/YAML taxonomy](chunkmaps_spec.html)
  [Agentsllms.txt entry map](https://chunkmaps.com/llms.txt)

Skill

Purpose

When to use

[Persuasion Flow](chunkmap_persuasion_flow.html)
[raw .md](skills/chunkmap-persuasion-flow.md)

Argument chains, counterarguments, CHUNKS_OK + Journey of Persuasion

Sales pitches, leadership, mediation, presentations, essays

[Agent Blackboard](skills/chunkmap-agent-blackboard.md)
[orchestrators guide](agent_orchestrators_chunkmaps.html)

Shared semantic blackboard + DAG prompt-trees

Multi-agent swarms that must not drift or invent conflicting state

[Policy to Pipeline](skills/chunkmap-policy-to-pipeline.md)
[enterprise guide](enterprise_architects_chunkmaps.html)

Compile unstructured policy into versioned, citable logic chunks

Enterprise AI pipelines, claims/exceptions, SOP to rules

## Skills README (source index)

Below is the current repo `skills/README.md`. Copy it into an agent context, or open the raw file in a new tab.

    [View Raw README](skills/README.md)

Copy Skills README

  # ChunkMaps Skills Library

Editable repository of **agent-usable skills**. Each `.md` file encapsulates one core ChunkMaps pattern so admins and agents can load, revise, and version method knowledge without hardcoding it into application code.

Hybrid positioning:

- **v1** - improved human communication & persuasion
- **Agentic Era** - semantic architecture for multi-agent systems

## How to use .md skill files

Start here: **[how-to-use-skill-files.md](./how-to-use-skill-files.md)**

Covers what a skill is, where the library lives, how to open skills from site CTAs, how agents/humans apply a skill step-by-step, and how to update or add skills. Formatted site pages: `chunkmap-skills.html` (library hub) and `how_to_use_skill_files.html`.

## Available skills

| Skill file | Purpose | When to use |
|---|---|---|
| [chunkmap-persuasion-flow.md](./chunkmap-persuasion-flow.md) ([TEMP guide](../TEMP/chunkmap_persuasion_flow.html)) | Structure argument chains, counterarguments, and persuasive narrative arcs (CHUNKS_OK + Journey of Persuasion). | Sales pitches, leadership comms, mediation, presentations, essays. |
| [chunkmap-agent-blackboard.md](./chunkmap-agent-blackboard.md) | Shared semantic blackboard + DAG prompt-trees for multi-agent swarms. | Orchestrating agents that must not drift or invent conflicting state. |
| [chunkmap-policy-to-pipeline.md](./chunkmap-policy-to-pipeline.md) | Compile unstructured policy prose into versioned, citable logic chunks. | Enterprise AI pipelines, claims/exceptions, SOP → machine-readable rules. |

## Raw paste-ready prompts

Full worked controller prompts (with USER INPUT examples) live in **[raw-prompts/](./raw-prompts/)**. They are regenerated from `_build_temp_site.py` on each TEMP rebuild. Guide pages expose a **Get Raw Prompt** button that opens the matching `.md` in a new tab.

## How to add a new skill

1. Copy an existing skill file as a template.
2. Name it `chunkmap-<pattern-slug>.md` (kebab-case).
3. Fill YAML frontmatter (`id`, `name`, `version`, `tags`).
4. Keep the standard sections: purpose, when to use, inputs, procedure, output schema, mini-example, guardrails, related links.
5. Add a row to the table above.
6. Bump `version` when you change behaviour of an existing skill (semver: patch = wording, minor = new steps/fields, major = incompatible output schema).

## Conventions

- **One pattern per file** - do not merge persuasion + blackboard into one skill.
- **Editable by non-engineers** - prefer markdown tables/outlines over code-only schemas.
- **Cite site articles** with relative links to `../TEMP/` (preview) or note the public HTML slug.
- **No secrets** - skills must not contain credentials or private customer data.
- Tone: authoritative, technical yet accessible.

## TEMP web preview

`_build_temp_site.py` copies this folder to `TEMP/skills/` on every rebuild so article skill CTAs resolve in the browser (`skills/<file>.md`). **Edit only the repo-root `skills/` files**-never treat `TEMP/skills/` as source of truth.

## Related site pages

- Skills hub: `chunkmap-skills.html`
- How-to: `how_to_use_skill_files.html`
- Persuasion: `chunkmap_persuasion_flow.html`, `best_persuasive_template.html`, `the_journey_of_persuasion.html`
- Agentic pillars: `agent_orchestrators_chunkmaps.html`, `enterprise_architects_chunkmaps.html`, `prompt_engineers_chunkmaps.html`

### Related

- [llms.txt](https://chunkmaps.com/llms.txt) (for LLM/agent discovery)

- [ai.txt](https://chunkmaps.com/ai.txt)

- [Raw prompts index](skills/raw-prompts/README.md)

- [How to use .md skill files](how_to_use_skill_files.html)

- [FAQ](faq.html)

- [ChunkMaps Spec (JSON/YAML)](chunkmaps_spec.html)

- [Agentic ChunkMaps hub](agentic_chunkmaps.html)

- [Articles & Sitemap](chunkmaps_articles.html)