How to Use These Guides

The Stunspot’s Guide to… repositories are Advanced Knowledge Bases: human-readable, but primarily designed as model-facing knowledge substrates.

Use them as:

1. Pick the right canon

Start from the Catalog. Choose the guide whose domain matches the task.

For cross-domain problems, combine guides deliberately. For example:

2. Prefer compiled packs first

Most guide repositories include:

knowledge-packs/
├── by-report/
├── compiled-packs/
└── omnibus/

For most AI systems, start with compiled-packs/.

Compiled packs usually give the best balance between:

3. Use by-report packs for precision

Use knowledge-packs/by-report/ when you need source-level separation, better citation discipline, or fine-grained retrieval.

This is usually best for:

4. Use omnibus packs for continuity

Use knowledge-packs/omnibus/ when the target system handles large single files well.

This is usually best for:

5. Cite the specific guide, not this gateway

This repository is the library index. Each guide owns its own citation metadata.

When citing a guide, use that guide’s own:

Use this gateway only when citing the catalog itself.

Common Pattern

A typical guide repository works like this:

Area Purpose
README.md Public orientation and quick-start explanation.
MANIFEST.md Human-readable structural inventory.
manifest.json Machine-readable structural inventory.
STATUS.md Release and completion status.
CITATION.cff Citation metadata.
LICENSE.md License terms for that guide.
docs/ Navigation, Pages support, and usage guidance.
knowledge-packs/by-report/ Individual source-report corpus.
knowledge-packs/compiled-packs/ Grouped upload packs.
knowledge-packs/omnibus/ Whole-corpus bundle.

Practical Loading Advice

When uploading guides into an AI workspace, avoid throwing everything into the same bucket without a reason. Choose a small set of relevant guides and tell the model what each one is for.

Example setup language:

Use Stunspot’s Guide to AI Systems as the primary doctrine for AI architecture, RAG design, agent boundaries, evals, and model operations.

Use Stunspot’s Guide to Semantics, Semiotics, and Symbols as the semantic hygiene layer for meaning, reference, ambiguity, symbols, interpretation, and citation discipline.

When the two guides overlap, treat AI Systems as the engineering layer and Semantics/Semiotics as the meaning-governance layer.

A public website landing page lives at:

https://stunspot.com/#guides