Knowledge Packs — Stunspot’s Guide to Sales Prospecting Strategy
The knowledge-packs/ directory contains the actual model-facing corpus. The docs/ directory is only navigation and usage guidance.
Use this page to choose the right upload format for AI Projects, RAG systems, NotebookLM-style tools, long-context workspaces, local search, or report-level editing.
Recommended Default
Use the compiled packs unless you have a specific reason not to.
They preserve the canon’s sequence while reducing file count from 16 individual reports to 5 grouped volumes. For most AI/RAG systems, this is the best balance between context coherence, upload ergonomics, and retrieval granularity.
Pack Types
| Pack Type | Files | Path | Best Use |
|---|---|---|---|
| Source reports | 16 | knowledge-packs/by-report/ |
Precise retrieval, selective upload, report-level citation, report editing, and narrow context loading. |
| Compiled packs | 5 | knowledge-packs/compiled-packs/ |
Recommended default: grouped coverage with lower file count and preserved canon sequence. |
| Omnibus | 1 | knowledge-packs/omnibus/ |
One-file import, local archive, full-corpus search, or strong long-context/RAG systems. |
Format Selection
Use source reports when you need precision
Choose knowledge-packs/by-report/ when the target system benefits from small retrieval units or when you need to cite, inspect, replace, or edit a specific report.
This is the right format for:
- report-level citation
- selective upload
- narrow RAG retrieval
- corpus maintenance
- differential updates
- targeted assistant behavior focused on one stage of the canon
Use compiled packs for most AI/RAG workflows
Choose knowledge-packs/compiled-packs/ when the system can ingest several medium-to-large files and you want the model to understand the sequence as a coherent doctrine.
This is the right format for:
- ChatGPT Projects
- Claude Projects
- NotebookLM-style tools
- custom RAG collections with moderate file-count limits
- strategy assistants
- sales operations assistants
- outbound planning workflows
- GTM diagnostic copilots
Use the omnibus when one file is better than many
Choose knowledge-packs/omnibus/ when the target tool works best with one file, or when you need a local archive of the complete corpus.
This is the right format for:
- one-file import
- offline archive
- local semantic search
- whole-corpus review
- strong long-context models
- systems where upload count is more constrained than file size
Avoid the omnibus in weak retrieval systems that degrade on very large single files. Bigger is not always wiser; sometimes it is just a whale in a kiddie pool.
Source Reports
Compiled Packs
| Volume | Compiled Pack |
|---|---|
| Vol. 1 A-D | Foundations of Commercial Reality and Signal |
| Vol. 2 E-G | Demand-Signal Intelligence and Opportunity Discovery |
| Vol. 3 H-K | Buyer Navigation, Commercial Influence, and Risk Mitigation |
| Vol. 4 L-N | Pipeline Systems, Outbound Operations, and Epistemic Transfer |
| Vol. 5 O-P | Diagnosis, Optimization, and Commercial Execution |
Omnibus
Directory Policy
docs/is navigation, GitHub Pages metadata, and usage guidance.knowledge-packs/by-report/contains the individual source reports.knowledge-packs/compiled-packs/contains grouped upload packs.knowledge-packs/omnibus/contains the whole-corpus bundle.- Do not create or reference
docs/reports/.