First archived public release of CogSecSkills — a defensive, harness-neutral library of 100 agentic tool-use skills for cognitive security and analytic tradecraft (Structured Analytic Techniques, source verification, deception detection, critical review), runnable identically under Claude Code, Codex, and Hermes.
Archived on Zenodo (permanent DOI):
- Concept DOI (all versions): https://doi.org/10.5281/zenodo.20804585
- Version DOI (v1.0.0): https://doi.org/10.5281/zenodo.20804586
This release
- 100 implemented skills across 7 taxonomy groups; machine-checkable multiharness conformance.
- Title cover now carries the archived DOI, an identity wordmark, and a live taxonomy band.
- Abstract rewritten as a single comprehensive plaintext paragraph.
- 622 tests passing at 91.18% coverage; validate + doctor clean.
The attached PDF is the rendered manuscript (also archived on Zenodo).