Learn-skills.dev content-memory

install

source · Clone the upstream repo

git clone https://github.com/NeverSight/learn-skills.dev

Claude Code · Install into ~/.claude/skills/

T=$(mktemp -d) && git clone --depth=1 https://github.com/NeverSight/learn-skills.dev "$T" && mkdir -p ~/.claude/skills && cp -r "$T/data/skills-md/agilebydesign/agilebydesign-skills/content-memory" ~/.claude/skills/neversight-learn-skills-dev-content-memory && rm -rf "$T"

manifest: data/skills-md/agilebydesign/agilebydesign-skills/content-memory/SKILL.md

source content

Content Memory Pipeline

Convert source documents to markdown (in place) and chunk them for agent memory. Content lives in

memory/

workspace/

. Chunks go directly into

memory/<domain>/<topic>/

Architecture

memory/ (project root): Content and chunks. No
```
converted/
```
or
```
chunked/
```
subfolders—chunks go directly in topic folders. Markdown in
```
<folder>/markdown/
```
.
workspace/ (optional): Source content for sync_and_chunk; copied to memory.
Convert to markdown/: PDF/DOCX → .md written in
```
<folder>/markdown/
```
for each folder.

Chunk: Reads from

source/<domain>/

memory/<domain>/

, writes to

memory/<domain>/<topic>/

When to Activate

"Add content to memory", "refresh memory", "ingest for agent"
"Sync workspace to memory", "convert and chunk"

Step 1: Convert to Markdown

Convert non-.md files to markdown in a

markdown/

subfolder per folder:

python scripts/convert_to_markdown.py --source <path> [--memory <domain>]
python scripts/convert_to_markdown.py --from source/CBE/domain_journeys_approach

Writes

.md

<folder>/markdown/

(e.g.

CB Domain/foo.pdf

→

CB Domain/markdown/foo.md

)

```
.md
```
files are skipped

Step 2: Chunk to Memory

Chunk markdown and write directly into memory topic folders:

python scripts/chunk_markdown.py --memory <domain> [--incremental]

Reads from:

source/<domain>/**/*.md

memory/<domain>/**/*.md

Writes to:
```
memory/<domain>/<topic>/
```
(no
```
chunked/
```
subfolder)
```
--incremental
```
: Only chunk new or modified files

Step 3: Sync Workspace (Convert + Copy + Chunk)

One command for workspace content:

python scripts/sync_and_chunk.py --workspace <topic> --memory <domain> [--incremental]

Converts non-.md to
```
<folder>/markdown/
```
(in workspace)

Copies

workspace/<topic>

→

memory/<domain>/<topic>

Chunks to
```
memory/<domain>/<topic>/
```

Chunking Strategy

Slide decks (
```

```
): One chunk per slide
Other docs (>200 lines): Split at
```
#
```
or
```
##
```
boundaries
Small files (<200 lines): Single chunk

Each chunk includes:

<!-- Source: path | file://url -->

Key Behaviors

Content in memory – Put content in
```
memory/<domain>/<topic>/
```
or sync from workspace.
Convert to markdown/ – Markdown written in
```
<folder>/markdown/
```
for each folder.
Chunks in memory – No
```
chunked/
```
subfolder; chunks go directly in
```
memory/<domain>/<topic>/
```
.
Incremental – Use
```
--incremental
```
to skip unchanged files.

Project-Specific Transformers

Location	Scope
`memory/<name>/transformers/`	Memory-specific
`.content-memory/transformers/`	Workspace-level

Each

.py

exports

EXTENSIONS

and

convert(path: Path) -> str

Scripts

Script	Purpose
`convert_to_markdown.py`	Convert to markdown/
`chunk_markdown.py`	Chunk to memory
`sync_and_chunk.py`	Convert + copy + chunk (workspace)

Run from workspace root. Set

CONTENT_MEMORY_ROOT

if needed.

Troubleshooting

Issue	Fix
No markdown	Run convert; then chunk. Or sync workspace to memory first.
Missing markitdown	`pip install "markitdown[all]"`